Clyde W. Oyster

Morphology of on-off direction-selective ganglion cells in the rabbit retina *

On-off direction-selective ganglion cells in rabbit retina have been stained by intracellular injection of horseradish peroxidase (HRP). The dendritic ramification is basically bistratified . Global asymmetries related to the preferred direction are not apparent, but the small diameter dendrites, spines, and complex branching pattern are consistent with models based on local, non-linear mechanisms for direction-selectivity.

Direction-selective retinal ganglion cells and control of optokinetic nystagmus in the rabbit

Abstract Approximately 25 per cent of the rabbits retinal ganglion cells are directionselective. These cells respond maximally to movement of a visual stimulus in a particular (“preferred”) direction. The class of direction-selective cells can be further subdivided into two groups; the “on-off” and “on-type” direction-selective cells. It has been suggested that the on-off cells may play a role in controlling the slow phase of optokinetic nystagmus. To test this hypothesis, we determined how the responses of directionselective cells, of both types, varied as a function of stimulus velocity. The stimulus was similar to those normally used to elicit nystagmus i.e. a continuous black and white striped grating. The velocity response curves for the direction-selective cells were compared to data on the open loop eye movement velocity of the slow phase of optokinetic nystagmus in rabbits. The results of this comparison indicate that the slow phase is primarily controlled by the on-type direction-selective cells with a secondary contribution from on-off cells at higher stimulus velocities.

Quantitative Morphology of Rabbit Retinal Ganglion Cells

Intraretinal (extracellular) injections of horseradish peroxidase were used to stain rabbit retinal ganglion cells. Five basic morphological ganglion cell classes were identified by quantitative analysis of dendritic branching patterns and computer reconstruction of dendritic ramification levels. Type 1 cells are characterized by a unistratified, radial dendritic morphology. The dendritic fields are of medium to large size. Subgroups ramify in either the outer or the inner part of the inner plexiform layer (i. p. l.). Type 2 cells have complex intricately branched dendritic morphologies with wide branch angles. They are comparable with type 1 cells in dendritic field size. Subgroups of this class include unistratified cells ramifying in the outer or inner part of the i. p. l. as well as cells with more complicated i. p. l. ramification schemes. Type 3 cells are somewhate similar to type 1 cells. A particular distinction is that they are much larger than type 1 cells at the same retinal eccentricity. Type 4 cells have a thin elliptical soma and a lobulate dendritic tree structure. Type 5 cells are a somewhat heterogeneous group with very small intricately branched dendritic fields. Since the number of anatomical groups is comparable with the number of physiological classes, it is tenable that the major physiological cell classes are associated with distinct dendritic morphologies.

Interplexiform cells in rabbit retina

Interplexiform cells, which may provide a centrifugal pathway within the retina, have been demonstrated in several vertebrate species. As revealed by Golgi stains, these cells are also present in rabbit retina and have several different morphologies. It appears that interplexiform cells are not only a general feature of the vertebrate retina, but are present in significant numbers.

Ganglion cell density in albino and pigmented rabbit retinas labeled with a ganglion cell-specific monoclonal antibody

Retinas from two rabbits, one normally pigmented and one albino, were labeled with monoclonal antibody AB5, which has been shown to be a specific marker for ganglion cells. This method obviates criteria for distinguishing among ganglion cells, displaced amacrine cells, and glia. Labeled cells were counted within small fields at some 2000 regularly spaced points on each retina. These counts were transformed to maps of ganglion cell density. In general, the density map for the pigmented retina was similar to those obtained by earlier studies with non-specific stains, thereby confirming the basic validity of most previous studies and demonstrating the applicability of AB5 labeling to work of this type. The ganglion cell density map of the albino retina was abnormal, showing a clear deficit of ganglion cells in the nasal portion of the visual streak. This result not only indicates that the albino anomaly has retinal effects, but also suggests a major impact on ganglion cells whose projections (in normal animals) are contralateral.

Information processing in the rabbit visual system

In a number of species, the visual information stream has become extensively edited at the retinal ganglion cell level. This is particularly true of the rabbits visual system, and many of the ganglion cells respond selectively to complex parameters of visual stimuli. Generally speaking, the response properties of the rabbits retinal ganglion cells show much the same degree of complexity as neurons in the visual cortex of higher vertebrates. An important difference, however, is that the synaptic connections in the retina have a known hierarchical pattern which makes the correlation of structure and function a more readily manageable task. In some cases, specific neural connections can be invoked to account for particular characteristics of the ganglion cell responses. The retinal information undergoes further alteration at each successive relay in the visual pathway. It is therefore tempting to suppose that cortical information processing mechanisms in higher vertebrates can be observed, say, two synaptic relays earlier in the rabbits visual pathway. Thus, investigation of the rabbit lateral geniculate nucleus may in some ~espects be comparable to studying second order cortical neurons in other animals. Another possibility is simply that transformations of visual information begin earlier in the rabbit, but occur in easier stages. In either case, investigation of information processing in the rabbit is facilitated by the development of complexity in discrete anatomical levels of the visual pathway.

Analysis of neuronal soma size distributions

Neuronal soma size distributions are often very skewed, leading to difficulty in their characterization and in the application of statistics commonly used to describe normally distributed variables. A particular family of curves, called Weibull functions, can be shown to fit cell size data extremely well. These functions not only characterize these skewed distributions, but do so in a way which permits powerful, sensitive comparisons of differences between distributions. Because of the flexibility of the Weibull functions, they may be used to describe a variety of morphological attributes. Several applications of these functions to neuroanatomical data are described using methods which require only graph paper and hand calculator for curve fitting. An extension of the basic curve fitting method permits mixtures of Weibull functions to be used in describing multimodal size histograms.

REFLEX EYE MOVEMENTS

&NA; Experimental studies of visually guided eye movements in the rabbit are used to develop the idea that human foveal tracking eye movements may be viewed as an image stabilizing system which uses visually gathered velocity signals to control the eye movements. These eye movements assist in keeping the image of a moving object on the fovea which is highly specialized for gathering and processing detailed visual information.