This paper reviews evidence that the superior colliculus (SC) of the midbrain represents visual direction and certain aspects of saccadic eye movements in the distribution of activity across a population of cells. Accurate and precise eye movements appear to be mediated, in part at least, by cells of the SC that have large sensory receptive fields and/or discharge in association with a range of saccades. This implies that visual points or saccade targets are represented by patches rather than points of activity in the SC. Perturbation of the pattern of collicular discharge by focal inactivation modifies saccade amplitude and direction in a way consistent with distributed coding. Several models have been advanced to explain how such a code might be implemented in the colliculus. Evidence related to these hypotheses is examined and continuing uncertainties are identified.

Selective labeling of intravitreal Procion yellow dye by presumed blue-sensitive cone photoreceptors has been demonstrated in primate retina. To determine whether Procion yellow is selective for this cone type in an unrelated vertebrate species, labeling by this dye was studied in goldfish retina, where cone pigment type can be directly inferred from photoreceptor morphology. At low vitreal concentrations of the dye (<0.4%), only cone outer segments were labeled. At vitreal concentrations of 0.4–0.5%, the inner segments of short-double cones and a subset of long single cones (presumed green-sensitive cones) were selectively stained. At still higher vitreal concentrations (0.6–0.7%), the inner segments of short-single cones and miniature short-single cones (presumed blue-sensitive cones) showed evidence of Procion label, but were not as heavily labeled. The inner segments of long-double cones and a subset of long-single cones (presumed red-sensitive cones) did not label at any of these concentrations. These results show that Procion yellow is not a selective marker for blue-sensitive cones in the goldfish retina. In addition, stained rod and cone nuclei were observed at each dye concentration, including those concentrations at which no inner segments were labeled.

Using the whole-cell patch clamp technique, we have examined the voltage-gated currents present in adult chicken cone cells. When calcium and calcium-gated currents are blocked, hyperpolarizing voltage steps elicit slowly increasing inward currents as has been shown for photoreceptors in other species. Unlike the case for other species, chicken cones appear to lack the inward-rectifying cationic current Ih that contributes to the shaping of the photovoltage. Instead of Ih, these cones possess an anionic inward-rectifying current that in kinetics, activation range and probably function is remarkably similar to Ih. This anion channel is unusual in that both nitrate and acetate are more permeant than chloride ions.

Evidence is presented that innate microstrabismus and abnormal cortical visual receptive-field properties can occur also in cats without any apparent involvement of the Siamese or albino genetic abnormalities in their visual system. A possible cause for microstrabismus in these cats may be sought in an abnormally large horizontal distance between blind spot and area centralis indicated by a temporal displacement of the most central receptive fields on both retinae.

Depth perception was found to be impaired in cats with innate microstrabismus. Behavioral measurements using a Y-maze revealed in four such cats that the performance in recognizing the nearer of two random-dot patterns did not improve when they were allowed to use both eyes instead of only one. The ability of microstrabismic cats to perceive depth under binocular viewing conditions only corresponded to the monocular performance of five normal cats.

Electrophysiological recordings were performed in the visual cortex (areas 17 and 18) of four awake cats, two normal, and two innate microstrabismic animals. Ocular dominance and orientation tuning of single neurons in area 17 and 18 were analyzed quantitatively.

The percentage of neurons in area 17 and 18 which could be activated through either eye was significantly reduced to 49.7% in the microstrabismic animals when compared to the normal cats (74.8%). “True binocular cells,” which can only be activated by simultaneous stimulation of both eyes, were significantly less frequent (1.6%) in microstrabismic cats than in normal animals (10.4%). However, subthreshold binocular interactions were identical in both groups of animals. In the strabismic animals, long-term binocular stimulation of monocular neurons did not give a clear indication of alternating use of one or the other eye.

The range of stimulus orientations leading to discharge rates above 50% of the maximal response, i.e. the half-width of the orientation tuning curves, was the same in the two groups of cats. However, orientation sensitivity, i.e. the alternation in discharge rate per degree change in stimulus orientation, was higher in cortical cells of normal cats than in those of microstrabismic cats.

In normal and microstrabismic cats, no clear sign of an “oblique effect,” i.e. the preference of cortical neurons for vertical and horizontal orientations compared to oblique orientations, could be found neither in the incidence of cells with horizontal or vertical preferred orientation nor in the sharpness of orientation tuning and sensitivity of these neurons.

In summary, the receptive-field properties reported here for awake innate microstrabismic cats are similar to those reported in the literature for anesthetized cats with varying degrees of albinism and for cats with artificial symmetrical strabismus surgically induced by sectioning the equivalent extraocular muscles in both eyes. Our innate microstrabismic cats may provide, however, an animal model for investigating the etiology of one form of naturally occurring strabismus.

Previous studies have reported that the surround responses of retinal ganglion cells weaken or disappear upon dark adaptation. The mechanism(s) by which this occurs is largely unknown, although changes in activity of retinal dopaminergic neurons have been implicated. In the light-adapted rabbit retina, the surround ON responses of OFF-center ganglion cells have been shown to be markedly reduced or abolished by a dopamine antagonist. This effect of a dopamine antagonist was recently shown to be reversed by the glycine antagonist strychnine and by compounds that elevate intracellular cAMP levels. The present study was conducted to determine whether strychnine and cAMP-elevating compounds could bring out the surround ON responses in OFF-center ganglion cells that are diminished upon dark adaptation. Extracellular recordings of OFF-center brisk ganglion cells were made from isolated, superfused retinal preparations. During the course of dark adaptation, the surround On responses of many cells decreased markedly.Application in both brisk-transient and brisk-sustained OFF-center ganglion cells. The center OFF responses of these cells, on the other hand, were not enhanced by strychnine. Of the cAMP-elevating compounds tested, 8-(4-chlorophenylthio) cyclic AMP was the most effective in bringing out the surround ON responses in dark-adapted OFF-center ganglion cells. The findings from this study suggest that under dark-adapted conditions glycinergic neurons inhibit the surround component of OFF-center ganglion cells. The release of glycine from these neurons is suggested to be regulated by a cAMP-dependent mechanism.

Much attention has been paid to the effect of various types of neonatally induced retinal insults on neurons of the cat lateral geniculate nucleus (LGN). Little is known about cellular adjustment to functional altertions commencing in the adult. The present study was aimed at examining the effect of monocular enucleation or retinal impulse blockade on mature neurons of the cat LGN and perigeniculate nucleus. In addition to labeling the relay neurons with cytochrome oxidase (CO) histochemistry and immunohistochemistry, and presumed interneurons with GABA immunohistochemistry, the two markers were also combined in the same section. The results showed that GABA-immunoreactive neurons in the LGN can be subdivided into two major groups: highly immunoreactive neurons (Hir) and moderately immunoreactive neurons (Mir). These two groups differed slightly in their size and CO levels. With monocular enucleation or TTX treatment, there was a reduction in the numerical density of Hir and a concomitant increase in Mir in the affected laminae. However, there was no evidence of a reduction in GABA immunoreactivity in neurons of the perigeniculate nucleus. With regard to relay cells our data were in agreement with our previous findings (Kageyama & Wong-Riley, 1985) that there was a statistically significant positive correlation between cell size and CO levels, so that neurons with large cross-sectional areas were predominately darkly reactive for CO (Dco), while medium-to-small neurons were mainly moderate to lightly reactive (M/Lco). After monocular enucleation or TTX injections, the numerical density and size of Dco were significantly reduced, while those of M/Lco were proportionally increased in the affected laminae, indicating a conversion of some Dco to M/Lco. The results indicate that the maintenance of CO levels in predominately the large relay cells and GABA levels in a subclass of small neurons of the mature cat LGN are activity-dependent and are highly sensitive to retinal insults.

Total RNA was purified from freshly isolated retinas of adult carp and injected into oocytes of Xenopus laevis (stage 5–6). Two to six days after injection, depolarizing voltage-clamp steps evoked a slowly activated outward currents as large as 3 μA. This current inactivated slowly with a single time constant (τ= 3.1 ± 0.24 S.E.M., for Vm= +30 mV). The current was inhibited by tetraethylammonium (3.8 mM for half-maximal inhibition). In the presence of Co2+ (1 mM) or barium methanesulfonate (40 mM), the current-voltage relationship shifted to slightly more depolarized values (5–10 mV); the maximal value of the current that was sensitive to Co2+ or Ba2+ treatments was only a small fraction (about 10%) of the TEA-sensitive current, and its current-voltage relationship was similar to that for uninjected oocytes. The reversal potential of the membrane current was studied with [K+]0 of 1–77 mM. For [K+]0 > 20 mM, the reversal potential changed with a slope of 63 mV (±;2 mV S.E.M.) per 10-fold change in [K+]0. The conductance was induced half-maximally at 17 mV (±;0.9 mV s.e.m.). The depolarization required for an e−fold increase in conductance was 13 mV (±;0.6 mV s.e.m.). From these results, we conclude that the injection of total RNA from carp retinas induces the formation of a membrane K+ channel in Xenopus oocytes. The channel formed has many of properties reported for the maintained outward current of goldfish horizontal and bipolar cells.

The inhibitory neurotransmitter gamma-aminobutyric acid (GABA) is found in the superior colliculus (SC) of many mammalian species. In cat, several distinct classes of putative GABAergic neuron have been identified using antibodies directed against the neurotransmitter. It is not known whether these classes are found in other species. To study this, we examined the distribution, morphology, ultrastructure, and synaptic organization of GABA immunoreactive neurons in the SC of the Rhesus monkey (Macaca mulatta). Antibody-labeled neurons were distributed throughout the monkey SC, but were most densely concentrated within the zonal and superficial gray layers (32.5% of the total). These neurons were all small cells ranging from 6.6–16.3 μm in average diameter, and had granule, pyriform, and horizontal morphologies. Four types of labeled profile were identified in single ultrathin sections with the electron microscope. Presynaptic dendrites (PSDs) contained pleomorphic vesicles, received synaptic input from unlabeled axon terminals, and sometimes formed symmeytric synaptic contacts with postsynaptic profiles. Two subtypes were found. One type contained loose accumulations of synaptic vesicles throughout the profile and had a distinctive varicose shape. The other type contained small discrete clusters of synaptic vesicles near the site of synaptic apposition. The former were much more common. Profiles with typical axon terminal morphology were also found. These profiles usually contained numerous flattened vesicles and formed symmetric synapses with postsynaptic profiles, both dendrites and cell bodies. Some conventional dendrites and myelinated axons were also labeled. Serial ultrathin section reconstructions revealed that PSDs formed complex synaptic relationships with other elements. Retinal terminals, identified by their characteristic pale mitochondria, established synaptic contacts with both types of PSD. These PSDs also established contact with each other, providing a possible anatomical substrate for disinhibition. We conclude that the monkey SC has multiple GABAergic cell types, similar to those found in cat may represent an organization common to both mammals and some other vertebrate species. The circuitry established by these cell types may provide a mechanism for disinhibition as well as inhibition in the mammalian SC.