As Cajal discovered in the late 19th century, the bird retina receives a substantial input from the brain. Approximately 10,000 fibers originating in a small midbrain nucleus, the isthmo-optic nucleus (ION), terminate in each retina. The input to the ION is chiefly from the optic tectum which, in the bird, is the primary recipient of retinal input. These neural elements constitute a closed loop, the centrifugal visual system (CVS), beginning and ending in the retina, that delivers positive feedback to active ganglion cells. Several features of the system are puzzling. All fibers from the ION terminate in the ventral retina and an unusual axon-bearing amacrine cell, the target cell, is the postsynaptic partner of these fibers. While the rest of the CVS is orderly and retinotopic, target cell axons project seemingly at random, mostly to distant parts of the retina. We review here the most significant features of the anatomy and physiology of the CVS with a view to understanding its function. We suggest that many of the facts about this system, including some that are otherwise difficult to explain, can be accommodated within the hypothesis that the images of shadows cast on the ground or on objects in the environment, initiate a rapid and parallel search of the sky for a possible aerial predator. If a predator is located, shadow and predator would be temporarily linked together and tracked by the CVS.

The bird visual system includes a substantial projection, of unknown function, from a midbrain nucleus to the contralateral retina. Every centrifugal, or efferent, neuron originating in the midbrain nucleus makes synaptic contact with the soma of a single unique amacrine cell, the target cell (TC). By labeling efferent neurons in the midbrain, we have been able to identify their terminals in retinal slices and make patch-clamp recordings from TCs. TCs generate Na+-based action potentials (APs) triggered by spontaneous EPSPs originating from multiple classes of presynaptic neurons. Exogenously applied glutamate elicited inward currents having the mixed pharmacology of NMDA, kainate, and inward rectifying AMPA receptors. Exogenously applied GABA elicited currents entirely suppressed by GABAzine and therefore mediated by GABAA receptors. Immunohistochemistry showed the vesicular glutamate transporter, vGluT2, to be present in the characteristic synaptic boutons of efferent terminals, whereas the GABA synthetic enzyme, GAD, was present in much smaller processes of intrinsic retinal neurons. Extracellular recording showed that exogenously applied GABA was directly excitatory to TCs and, consistent with this, NKCC, the Cl− transporter often associated with excitatory GABAergic synapses, was identified in TCs by antibody staining. The presence of excitatory retinal input to TCs implies that TCs are not merely slaves to their midbrain input; instead, their output reflects local retinal activity and descending input from the midbrain.

An avian retinopetal nucleus, the isthmo-optic nucleus (ION), is known to receive predominant inputs from the ipsilateral optic tectum. We injected biocytin into the ION in the Japanese quail, and retrogradely labeled tectal neurons projecting to the isthmo-optic (IO) neurons, or the tecto-IO neurons, with an extraordinary Golgi-like quality. Somata of the tecto-IO neurons were located in layer 9 of the tectum. The tecto-IO neurons did not have apical dendrites extending into superficial retino-recipient layers (layers 2–7), but had descending dendrites ramified in layers 9–12. They also possessed short ascending dendrites ramifying in the upper half of layer 9. This dendritic morphology suggests that main input to the tecto-IO neurons may not be of retinal origin. The tecto-IO neurons were spatially arranged in a regular pattern. Distances between neighboring tecto-IO neurons were 50–100 μm. The dendrites of each tecto-IO neuron were not widely dispersed in the horizontal plane, and were confined in a vertically oriented column of 100–200 μm diameter. They possessed axon collaterals extending horizontally in layers 12 and 13. The estimated total number of the tecto-IO neurons was approximately 7000–10,000, which is almost identical to the total cell number of the IO neurons. To label a small number of the tecto-IO terminals, biocytin was injected into a confined area of the optic tectum. The tecto-IO fibers densely arborized in a restricted space of the ION, which is comparable to the dimension of dendritic arborization of individual IO neurons. It is suggested that single tecto-IO neurons may make contact with single IO neurons. IO neurons are known to make synaptic contact with single target cells (association cells of Cajal) in the retina (Uchiyama & Ito, 1993; Uchiyama et al., 1995). The arborization pattern of the tecto-IO neurons' dendrites indicates that the tecto-IO neurons receive very local information in sensory and sensorimotor coordinate space. The morphology of the tecto-IO terminals suggests that the spatially confined information that drives the tecto-IO neurons is sent in parallel to single IO neurons, and then further to single retinal association cells of Cajal. Cellular-level restricted projection patterns of the tecto-isthmo-retinal system may reflect its function for centrifugal control of retinal function.

The effects of bilateral lesions of the centrifugal visual system (CVS) on the visual-discrimination capacity were studied in pigeons. Three different behavioral experiments, each testing different aspects of visual analysis, were performed. In the first two experiments, a grain-grit discrimination task and a visual-acuity determination, stimuli were presented in the frontal binocular visual field. A third experiment investigated the early detection of slow moving objects, introduced into the monocular lateral visual field. After bilateral lesions in the nucleus isthmo-opticus (ION) and in the ectopic nucleus isthmo-opticus (EION), a multiple linear regression analysis was employed to correlate the postoperative performance in all three tasks with the amount of structure loss within ION and EION. Deficits in the grain-grit discrimination procedure were a function of the ION lesion extent and did not depend on EION damage. Thus, these two structures could be functionally differentiated for the first time. Neither the ION nor the EION seems to be involved in visual- acuity performance or the early detection of large shadows moving forward through the visual field. Our data support the hypothesis that the CVS is involved in pecking and food selection among static stimuli at a short viewing distance in ground-feeding birds such as pigeons and chickens.

Two types of centrifugal pathways to the retina have been found in the vertebrates, according to the location of the cell bodies and presence or absence of connections with the optic tectum. One type is represented by the isthmo-optic nucleus (ION) of birds and, therefore, termed “ION-type” retinopetal system. The other type is termed “non-ION-type” retinopetal system. The ION-type retinopetal systems have been found in the cyclostomes, teleosts, reptiles, and birds. This review describes the anatomy and physiology of the ION-type retinopetal systems, mainly of birds and teleosts. On the basis of anatomical and physiological evidence cited in this review, the ION-type retinopetal systems can be regarded as the tectofugal pathways to the retina. The function of the ION-type retinopetal systems is discussed in detail, with special emphasis on their relation to the role of the tectum in mediating visuomotor behavior.

In birds, neurons of the isthmo-optic nucleus (ION), as well as “ectopic” neurons, send axons to the retina, where they synapse on cells in the inner nuclear layer (INL). Previous work has shown that centrifugal axons can be divided into two anatomically distinct types depending on their mode of termination: either “convergent” or “divergent” (Ramon y Cajal, 1889; Maturana & Frenk, 1965). We show that cytochrome-oxidase histochemistry specifically labels “convergent” centrifugal axons and target neurons which appear to be amacrine cells, as well as three “types” of ganglion cells: two types found in the INL (displaced ganglion cells) and one in the ganglion cell layer. Labeled target amacrine cells have distinct darkly labeled “nests” of boutons enveloping the somas, are associated with labeled centrifugal fibers, and are confined to central retina. Lesions of the isthmo-optic tract abolish the cytochrome-oxidase labeling in the centrifugal axons and in the target amacrine cells but not in the ganglion cells. Cytochrome-oxidase-labeled ganglion cells in the INL are large; one type is oval and similar to the classical displaced ganglion cells of Dogiel, which have been reported to receive centrifugal input; the other type is rounder. Rhodamine beads injected into the accessory optic system results in retrograde label in both types of cells, showing that two distinct types of displaced ganglion cells project to the accessory optic system in chickens. The ganglion cells in the ganglion cell layer that label for cytochrome oxidase also project to the accessory optic system. These have proximal dendrites that ramify in the outer inner plexiform layer. Neither the target amacrine cells nor either of the displaced ganglion cells are immunoreactive for the inhibitory transmitter gamma aminobutyric acid. At least some of the target amacrine cells may, however, be cholinoceptive: we found that the antibody to the alpha-7 subunit of the nicotinic ACh receptor labels a population of cells in the INL that are similar in location, size, and the presence of labeled bouton-like structures to those we find labeled with cytochrome oxidase. This antibody also labels neurons in the ION proper but not ectopic cells. In conclusion, it appears that cytochrome oxidase may be a marker for “convergent” centrifugal axons and at least one of their target cells in the INL.

This commentary extends the review by Hiroyuki Uchiyama (1989) entitled “Centrifugal pathways to the retina: influence of the optic tectum” Visual Neuroscience 3, 183–206. A further scrutiny is made of the topography of the isthmo-retinal projection of the pigeon, which provides the most fully investigated example of a retinopetal projection. It is suggested that the topographical organization in this system is compatible with it's input being derived from the lower visual field (and upper retina) and it's output being directed to the horizon and lower retina. This is another form of attentional hypothesis for the centrifugal pathway, compatible with a role in ground feeding, and is testable by current tracing techniques in neurobiology.

The isthmo-optic nucleus (ION) is the main source of efferents to the retina in birds. Isthmo-optic neurons project in topographical order on amacrine cells in the ventral parts of the retina, and a subclass of these known as proprioretinal neurons project onto the dorsal retina. We propose that, through the intermediary of the amacrine target cells, activity in the isthmo-optic pathway excites ganglion cells locally in the ventral retina but inhibits those in dorsal regions. This circuit would thereby mediate centrifugally controlled switches in attention between the dorsal retina, involved in feeding, and the more ventral parts, involved in scanning for predators. This hypothesis accounts for a wide range of disparate data from behavior, comparative anatomy, endocrinology, hodology, and neurophysiology.

The visual system of birds includes an efferent projection from a visual area, the isthmo-optic nucleus in the midbrain, back to the retina. Using a combination of anterograde labeling of efferent fibers, reconstruction of dye-filled neurons, NADPH-diaphorase staining, and transmission electron microscopy, we have examined the distribution of efferent fibers and their synaptic structures in the chicken retina. We show that efferent fibers terminate strictly within the ventral retina. In two completely mapped retinas, only 2 fibers from a total of 15,359 terminated in the dorsal retina. The major synapse made by each efferent fiber is with a single efferent target amacrine cell (TC). This synapse consists of 5–25 boutons of 2 μm diameter, each with multiple active zones, pressed into the TC soma or synapsing with a basketwork of rudimentary TC dendrites in the inner nuclear layer (INL). This basketwork, which is sheathed by Muller cell processes, defines a private neuropil in the INL within which TCs were also seen to receive input from retinal neurons. In addition to the major synapse, efferent fibers typically produce several very thin processes that terminate nearby in single small boutons and for which the soma of a local amacrine cell is one of the likely postsynaptic partners. A minority of efferent fibers also give rise to a thicker process, terminating in a strongly diaphorase-positive ball about 5 μm in diameter.

In 1895, by means of the Golgi method, Santiago Ramón y Cajal discovered a cell having a unique morphology in the avian retina. This cell had its cell body in the amacrine cell level of the inner nuclear layer, only a few rudimentary dendrites at the outermost level of the inner plexiform layer (IPL), and a long axon coursing horizontally and terminating in the IPL. Despite having defined amacrine cells as cells without axons, Cajal named this cell type “association amacrine cell” (AAC). This discovery was not confirmed by other investigators for nearly a century. Very recently, however, isthmo-optic target cells (IOTCs), which receive the terminals of centrifugal fibers emanating from the isthmo-optic nucleus, have been identified as one type of AAC. As summarized and discussed in this review, the morphology of the AACs as described by Cajal has been completely confirmed. However, since these cells appear to be classical polarized, monoaxonal neurons and lack the dendritic interactions that are typical of amacrine cells, they should be regarded as a distinct type of retinal interneuron and not as amacrine cells.

Birds have a well-developed centrifugal pathway from the isthmo-optic nucleus to the retina. In the present study, receptive-field properties of the isthmo-optic (IO) neurons and suppressive effect of remote stimuli far beyond the “classical” receptive fields were examined electrophysiologically in the Japanese quail. Receptive fields emitting more than half the maximal ON–OFF response to a small spot of light measured 4.3 ± 1.9 deg (n = 37) in diameter. A stationary spot of light was presented at a remote point (35–76 deg) away from the receptive-field center, besides a stimulus for the receptive-field center, with various onset time difference. The peripheral spot, when turned on 50–100 ms before the center spot, maximally suppressed the ON and OFF responses to the center spot. In most of the IO neurons examined (13 out of 15 units), even small remote stimuli (2–5 deg in diameter) significantly suppressed the ON–OFF responses to the center stimulus. The suppressive effects of remote stimuli were seen to be extended to, at least, two thirds of the entire visual field of a single eye. Such wide suppressive fields indicate that the IO neurons may compete with each other for activity in a very long-range scale. Simulation using a simple static model based on three basic principles of processing of the centrifugal visual system— topographical input, long-range competition, and local gain enhancement—suggests that the system could function as an autonomous mechanism for attentional object selection.

Centrifugal modulation of visual responsiveness of tectal cells by the isthmo-optic nucleus (ION) through the retina was studied in homing pigeons. Visual activity evoked by computer-generated stimuli was reduced by an average of 59% in tectal cells whose receptive fields (RFs) either overlapped with, or were close to, those of isthmo-optic cells whose activity was blocked by the injection of lidocaine through micropipettes. Activity usually recovered to 87% of pre-drug controls in 8–17 min (average 12.3 min) after stopping lidocaine injections. Those tectal cells whose RFs were far from those of ION cells did not show clear-cut changes in their visual responsiveness to isthmo-optic lidocaine application. The spatial relationship between receptive fields of tectal and isthmo-optic cells, saline controls, as well as the specificity, reproducibility and reversibility of effects of ION-injected lidocaine on tectal activity, show that this chemical action is pharmacological, not toxicological. Neuronal circuitry underlying centrifugal modulation of tectal activity by isthmo-optic cells is discussed.