Michael W. Levine

The relative capabilities of the upper and lower visual hemifields

Visual performance is better in the lower visual hemifield than in the upper field for many classes of stimuli. The origin of this difference is unclear. One theory associates it with finer-grained attention in the lower field, an idea consistent with a change in relative efficacy with task difficulty. The first experiment in this study confirmed a lower hemifield advantage for discriminating a range of stimuli, including those that differ in contrast, hue, and motion. An identical paradigm revealed an upper field advantage when stimuli differed in their apparent distances from the observer. Presentations of stimuli in the upper or lower hemifield were interlaced to reduce the likelihood of possible artifacts or biases. A second experiment varied the difficulty of these discriminations, showing that difficulty does not determine field preference. Thus, an attentional mechanism is not a likely explanation for these preferences.

The distribution of the intervals between neural impulses in the maintained discharges of retinal ganglion cells

Simulated neural impulse trains were generated by a digital realization of the integrate-and-fire model. The variability in these impulse trains had as its origin a random noise of specified distribution. Three different distributions were used: the normal (Gaussian) distribution (no skew, normokurtic), a first-order gamma distribution (positive skew, leptokurtic), and a uniform distribution (no skew, platykurtic). Despite these differences in the distribution of the variability, the distributions of the intervals between impulses were nearly indistinguishable. These inter-impulse distributions were better fit with a hyperbolic gamma distribution than a hyperbolic normal distribution, although one might expect a better approximation for normally distributed inverse intervals. Consideration of why the inter-impulse distribution is independent of the distribution of the causative noise suggests two putative interval distributions that do not depend on the assumed noise distribution: the log normal distribution, which is predicated on the assumption that long intervals occur with the joint probability of small input values, and the random walk equation, which is the diffusion equation applied to a random walk model of the impulse generating process. Either of these equations provides a more satisfactory fit to the simulated impulse trains than the hyperbolic normal or hyperbolic gamma distributions. These equations also provide better fits to impulse trains derived from the maintained discharges of ganglion cells in the retinae of cats or goldfish. It is noted that both equations are free from the constraint that the coefficient of variation (CV) have a maximum of unity. The concluding discussion argues against the random walk equation because it embodies a constraint that is not valid, and because it implies specific parameters that may be spurious.

Variability in ganglion cell firing patterns; implications for separate “on” and “off” processes

Abstract Correlation matrices and autocorrelation functions were calculated from responses of goldfish retinal ganglion cells to repeated photic stimuli effective only for the long-wavelength cones. Both of these analyses indicated that the ON response is generated by a process independent of the one that generates the OFF response, while the maintained discharge is a mixture of the two. In most cases, the maintained discharge seems to be primarily a product of the process responsible for the OFF response. The properties of the variability of the center and surround mechanisms were compared and found to be distinct: thus the center and surround of the receptive field each possesses a separate ON and OFF process. The transient characteristics of the responses are generated primarily within these processes. A model is presented to account for these findings.

Magnocellular and parvocellular visual pathway contributions to visual field anisotropies

It is well established that sensitivity is not necessarily equivalent at isoeccentric locations across the visual field. The focus of this study was a psychophysical examination of the spatial sensitivity differences between the upper and lower visual hemifields under conditions biased toward the presumed magnocellular or parvocellular visual pathway. Experiment 1 showed higher contrast sensitivity in the lower visual field when visual sensitivity was biased toward the parvocellular pathway; no visual field anisotropy was found when sensitivity was biased toward the magnocellular pathway. Experiment 2 showed that the magnitude of the contrast sensitivity anisotropy within the presumed parvocellular pathway increased when test targets of higher spatial frequency were used. The results of this study have relevance for the design both of psychophysical paradigms and clinical training programs for patients with heterogeneous visual field loss.

X-like and not X-like cells in goldfish retina

Cat retinal ganglion ceils can be divided into three functionally distinct classes according to the way they process spatial information. X-cells respond as if they sum linearly inputs to different areas of their receptive fields. while Yand W-cells show marked non-linear properties (Enroth-Cugeil and Robson. 1966; Stone and Fukuda. 1974). Although numerous other physioIogicaf (e.g.. Enroth-Cugeif and Robson, 1966; Hoffman, 1973; Stone and Hoffman, 1972) and anatomical (Cleland, Levick and W&ssle, 1975: Fukuda and Stone, 1975) properties are correlated with this classification system, the defining characteristic for X-, Y-. and W-cells is based on the way they respond to spatially distributed inputs. The dichotomy of linear and nonlinear ganglion cells has been extended to other mammals (DeMonasterio, Gouras and Tolhurst, 1976; Sherman. Wilson. Kaas and Webb, 1976). and a similar division can be made in at least one cold blooded animal. the eel (Shapley and Gordon, 1978). In this study. we extend this classification to ganglion cells in the goldfish retina.

A psychophysical demonstration of goldfish trichromacy

The goldfish retina has been used extensively as a convenient preparation for studying the electrophysiology of trichromatic color vision. However, the evidence underlying the assumption of goldfish trichromacy is not strong enough to warrant its uncritical acceptance. While tri~hromacy at the level of the receptor has been well established (Marks, 1965; Tomita, Kaneko. Murikami and Pautler, 1967). Spekreijse, Wagner and Wolbarsht (1972) found that when the red and blue chromatic mechanisms are both present in the same goldfish ganglion cell, they are always synergistic. If red and blue inputs are always paired, then they act as a single chromatic mechanism and the fish is a dichromat. In contrast to Spekreijse er (pi., Beauchamp and Lovasik (1973) found that all optic nerve fibers in intact goldfish that respond to blue light respond in the same manner to green light, and in an opposite manner to red. This contradiction of the findings of Spekreijse et al. has not yet been explained satisfactorily: however, since two chromatic mechanisms are always paired in the Beauchamp and Lovasik results, they also suggest the possibility that goldfish are dichromats. Behavioral studies have not resolved the question of whether goldfish have a trichromatic visual system. McCleary and Bernstein (1959) demonstrate that goldfish are at Ieast dichromats by training them to discriminate a green light from a red light, with brightness cues removed. Cronly-Dillon and lMuntz (1965) obtained a spectral sensitivity function for the goldfish. using the fish’s optokinetic response. The sensitivity function had three distinct peaks, perhaps indicating the existence of three chromatic mechanisms but certainly not implying behavioral trichromacy. Yager (1967) used an operant task to measure the spectral saturation function of the goldfish. He found that no wavelength was confused with white light, providing evidence that goldfish do in fact have a t~chromatic visual system. However, Yager spanned the visible spectrum with only 12 test wavelengths and did not carefully explore the regions surrounding the two local minima, 490 and 600 run. It is therefore quite possible that there remains an untested wavelength the goldfish cannot discriminate from white light. Yager (1974) recently reported measurements of ~tu~tion functions under different chromatic adap tation conditions. Although no confusion point was found, test wavelengths vvere separated by 2Onm. so these data are no more conclusive than those from the previous study. The minimum around 600 nm is probably more significant than the one at 490nm. If goldfish ganglion cells are organized in either of the configurations found by Spekreijse et u[., or Beauchamp and Lovasik, the major opponency is between the red and the green cone mechanisms. The wavelength for which this opponent system is n&led will be the one for which saturation goes to zero. and should therefore be confused with white; this null point must lie above 530nm, the peak of the green pigment. The purpose of the present experiment was to resotve the question of possible functional dichromacy through a careful investigation of the goidfish’s saturation discrimination abilities, in the region of the spectrum where Yager found his strongest minimum. A spectral sensitivity function was also required, in order to scale correctly the intensities of the spectral lights when comparing them to white light.

Tailoring of variability in the lateral geniculate nucleus of the cat

Abstract. Variability is usually considered an unwanted component in a sensory signal, yet the visual system does not seem to filter out the noise. On the contrary, noise is ‘tailored’ to scale with the signal size. We show that this tailoring occurs in the lateral geniculate nucleus, preferentially in X-cells, which are the cells most likely to transmit pattern information. Tailoring the variability to the signal size may be the visual system’s way of providing the right amount of variability for a signal of any magnitude at all times during the computation.

Variability of responses of cat retinal ganglion cells

Previous studies of the variability of firing of retinal ganglion cells have led to apparently contradictory conclusions. To a first approximation, the variance of rate of maintained discharges of ganglion cells in cat is independent of the mean firing rate. On the other hand, the variability of responses to abrupt changes in lighting of ganglion cells in goldfish increases with increasing firing rate. To examine whether the difference is due to differences between species, we examined the variability of responses of cat ganglion cells, and find it similar to that of goldfish ganglion cells. The variance of rate of ganglion cells is neither independent of mean rate, as might be expected from maintained discharges, nor directly proportional to the mean rate, as it is for cat cortical cells. Rather, there is a nonlinear relationship between variance of rate and mean rate.

A comparison of properties of goldfish retinal ganglion cells as a function of lighting conditions during dissection.

Abstract Recordings of action potentials were made from single ganglion cells in the isolated retinas of goldfish; retinas were dissected from the eye either in normal room lights or under dim red illumination. Cells from retinas isolated in normal lighting were usually not spectrally opponent, having a sensitivity dominated by the long wavelength sensitive cones; rod input to these cells could not be demonstrated. Cells from retinas isolated in dim red light, on the other hand, commonly were spectrally opponent and generally had demonstrable rod input. Other properties of the cells from retinas isolated under either lighting condition were the same.

Variability in the maintained discharges of retinal ganglion cells

The statistics of maintained discharges of goldfish and cat retinal ganglion cells are examined, and simulations are made with integrate-and-fire models. Actual interval distributions are best fitted with a hyperbolic gamma distribution; however, interval distributions derived from simulations demonstrate that this need not imply a skewed variable signal. The high-pass filtering in the maintained discharges of ganglion cells is weaker than would be expected from the transfer function for photic signals; this phenomenon, as well as the observation of two peaks in some of the interval distributions, is modeled by transient changes in the firing threshold after each impulse. Increases in firing rate owing to changes in the level of background illumination are accompanied by declines in the coefficient of variation; simulations predict an inverse-square-root relationship between the rate and the coefficient of variation, an approximation that may be improved by again assuming changes in the firing threshold.