Lateral interactions in the control of visual sensitivity

Effect of eccentricity on luminance-pedestal flicker thresholds

We investigated the effect that spatially coincident luminance increments (luminance pedestals) have on flicker thresholds at several eccentricities and target sizes. Luminance pedestals elevated flicker amplitude-thresholds more when stimuli were presented eccentrically, both at low (4 Hz) and high (20 Hz) temporal frequencies. Altering the size of the eccentric stimulus failed to equate central and eccentric thresholds at all pedestal amplitudes. Comparisons with flicker thresholds at various background luminances suggests that the increase in luminance-pedestal flicker thresholds peripherally is due to increased suppressive rod-cone interactions, increased effectiveness of luminous contrast on edge-sensitive flicker mechanisms, as well as increased gain in the light adaptation response.

Interactions between flicker thresholds and luminance pedestals

We investigated the interactions between flicker thresholds and luminance pedestals using threshold versus contrast (TvC) and method of constant stimuli paradigms. High amplitude luminance pedestals were found to elevate flicker thresholds, but low amplitude luminance pedestals were unable to reduce flicker thresholds. Luminance pedestals elevated flicker thresholds more at low temporal frequencies. A simple model based on local light adaptation was able to capture the general form of the TvC functions. Our results suggest that flicker thresholds derived in the presence of a luminance pedestal (luminance-pedestal flicker) may vary from those obtained by modulating about a mean luminance (mean-modulated flicker).

Saturation revealed by clamping the gain of the retinal light response

The saturation nonlinearity of the retinal light response in human was measured by a psychophysical technique in which the adaptive gain control mechanism was clamped by the presence of a fixed surround in a small (7') foveal test field. Gain clamping was established by showing that the normal variation in temporal summation properties with test intensity was abolished in the gain clamping paradigm. The static saturation function constructed from the increment/decrement asymmetries around a range of base intensities was shown to conform more closely to the Naka-Rushton hyperbolic saturation equation than to three other candidate nonlinearities.

Differences in adaptation between on- and off-centre ganglion cells and rod-mediated cone sensitization in cat retina

1. Response properties of on- and off-centre retinal ganglion cells were investigated in cats. The stimulus parameters were selected so as to demonstrate interactions between the rod and the cone systems. 2. Response versus log stimulus intensity (R-log I) functions were determined for the receptive field centres while both test stimulus irradiance and the background illumination were varied over a range up to 7 log units. In order to determine the course of adaptation to chromatic stimuli, threshold versus intensity (t.v.i.) functions were measured over a wide range of adaptation levels. 3. An increase in background illuminance produced a shift of the R-log I functions to higher irradiances of test stimuli in most ganglion cells, indicating a desensitization of the centre response in the presence of background lights. Using test stimuli which most efficiently stimulate the rods (501 nm), clear differences could be seen in the adaptation behaviour of on- and off-centre ganglion cells. Chromatic backgrounds (blue-green and orange) reduced the responses of off-centre cells more than those of on-centre cells (the difference between them amounting to as much as 2 log units). Simultaneously, equivalent t.v.i. functions had significantly steeper slopes (0.94 and 1.1) in the linear proportions of off-centre cells compared to on-centre cells (0.76 and 0.75) under light levels mediated by rods. Such differences were not observed when a test stimulus of 575 nm was used which resulted primarily in stimulating the long-wavelength cone (L-cone) system. 4. In a subpopulation of off-centre cells (20% of the total number of off-centre cells recorded), a strikingly different adaptation behaviour was observed. Here, the presentation of a dim short-wavelength background produced a shift of R-log I functions to lower test stimulus irradiances. The receptive field centre became even more sensitive, by up to 1.5 log units, in the presence of dim adapting backgrounds rather than in the dark-adapted state. Accordingly, the t.v.i. function did not increase monotonically but showed a 'dip' in the presence of dim backgrounds. Only at photoic levels, the t.v.i. functions revealed a response behaviour similar to the other ganglion cells. The sensitization with dim backgrounds was only observed in the case of test stimuli designed to stimulate the cone system (575 nm) and in the presence of a rod-adapting blue-green background.(ABSTRACT TRUNCATED AT 400 WORDS)

Visual sensitivity to spatially sampled modulation in human observers

Thresholds were measured for detecting spatial luminance modulation in regular lattices of visually discrete dots. Thresholds for modulation of a lattice are generally higher than the corresponding threshold for modulation of a continuous field, and the size of the threshold elevation, which depends on the spacing of the lattice elements, can be as large as a one log unit. The largest threshold elevations are seen when the sample spacing is 12 min arc or greater. These results are similar to those observed by Burr, Ross and Morrone [Vision Research, 25, 717-727 (1985)], who proposed an explanation based on a compressive point nonlinearity. Although their explanation is not consistent with the present data, the results may be explained in terms of nonlinear saturation of a spatially opponent stage early in the visual pathway. Theories based on response compression cannot explain the further observation that the threshold elevations due to spatial sampling are also dependent on modulation frequency: the greatest elevations occur with higher modulation frequencies. The idea that this is due to masking of the modulation frequency by the spatial frequencies in the sampling lattice is considered.

Spatial sensitization in the short wavelength sensitive pathways under dichoptic viewing conditions

Spatial sensitization (Westheimer) functions were measured under conditions that isolated the short wavelength sensitive pathways. The variable diameter pedestal and the test probe were either presented to the same eye (monoptic presentation) or to different eyes (dichoptic presentation). The most significant new finding was that a dichoptically presented, small, blue pedestal caused threshold elevations of about 1-2 long units for an S cone detected probe. However, a large pedestal caused little or no change in threshold. This result contrasts with previous results using white light stimuli, which showed that steadily presented dichoptic pedestals caused little or no threshold change. Furthermore, we show there is little masking when the probe is detected by the isolated M or L cone pathways. These data thus demonstrate a binocular, size dependent interaction revealed only when S cones detect the probe.

The role of spatial filtering in sensitivity regulation

The role of spatial filtering in controlling sensitivity to increments is hard to evaluate under normal viewing conditions because eye movements lead to a confounding of spatial and temporal transients. We measured sensitivity to increments on different sized backgrounds in photopic and scotopic vision when the backgrounds were stabilized on the retina, thus eliminating temporal transients. The saturating effect of small fields on photopic thresholds was preserved under these conditions indicating that spatial filtering by retinal cells is critical in maintaining photopic sensitivity. Some effect of spatial pattern on sensitivity in stabilized vision was also observed in scotopic vision, although it was much smaller than was observed in photopic vision. The interaction effects between rod and cone systems that are observed with small backgrounds were also preserved in stabilized vision, implicating a very peripheral site for the generation of these interactions.

Spatial sensitization of the B cone pathways

Spatial sensitization functions were determined under chromatic adaptation conditions designed to isolate the short-wavelength-sensitive pathways (B cone pathways). Spatial sensitization occurs over large retinal dimensions for B cone detected targets and can be produced both by signals from B cones and by signals from R and/or G cones. Field sensitivity measures demonstrate that the mechanism producing sensitization on short-wavelength backgrounds has the spectral sensitivity of B cones. The field sensitivity on long-wavelength backgrounds does not correspond to any receptor spectral sensitivity or any pi mechanism, but suggests that opponent interaction between medium and long-wavelength receptors produces sensitization of B cone detected targets.

Effect of surround configuration on increment thresholds of a tiny violet flash

Increment thresholds were measured for a 5', 435-nm test flash (50 or 200 msec) on background fields of various wavelengths and spatial configurations. As the field diameter was increased, the detection threshold for the 50-msec test decreased rapidly and leveled off when the field diameter exceeded 20'. For short-to-middle wavelength fields (much less than or equal to 503 nm), the detection threshold for the 200-msec test varied with field diameter in a similar manner to those observed for the 50-msec tests, while for middle-to-long wavelength fields (much greater than or equal to 544 nm), the detection threshold continued to decrease even when the field diameter exceeded 20'. The temporal-integration properties and the threshold vs intensity curves were determined for 465-nm and 579-nm fields to specify the mechanisms responsible for the detection. It was shown that the detection thresholds of the test upon the large 579-nm field were characterized by the short-wavelength mechanism isolated by an intense yellow adapting field, whereas the detection thresholds of the test upon the small 579-nm or the 465-nm field showed an intrusion of a more sensitive alternate mechanism, presumably the middle-wavelength mechanism. It was suggested that the sensitivity of the S cone signals might be affected by spatially distributed signals from the longer-wavelength mechanisms.

The spectral sensitivities of the middle- and long-wavelength cones: an extension of the two-colour threshold technique of W S Stiles

When a tiny centred test flash is presented on a small concentric background, the threshold rises with background radiance more quickly than Weber's law would predict. It is argued that under such conditions it is possible, by means of a test sensitivity method, to isolate either the M-cone or the L-cone types throughout the visible spectrum. As predicted, double-branched M- and L-cone tyr functions are found when the test flash and the field are of the same wavelength. From the independent vertical displacements of the two branches as test wavelength is varied, it is possible to derive spectral sensitivities that agree well with dichromatic sensitivities and König fundamentals. The test sensitivities deviate from pi 4 at longer wavelengths and from pi 5 at shorter wavelengths.

Background size and saturation of the rod system

To determine the effect of background size on saturation of the human rod system, we measured threshold-vs-intensity (tvi) functions for a 24'-diameter, middle-wavelength test stimulus against long-wavelength background fields that were either 2 degrees or 8 degrees in diameter. When thresholds were measured against a series of background intensities presented during a single experimental session (a standard tvi procedure), rod increment thresholds were higher against the smaller background and eventually disappeared above cone increment thresholds at moderate background intensities. This result suggests that background diameter does influence rod system saturation. However, when adaptation to the smaller background was limited to a 5 min exposure to a single background intensity per experimental session, then thresholds against that background were rod- rather than cone-mediated and were comparable to rod thresholds against an 8 degrees background. Therefore, under these stimulus conditions, background size apparently does not influence rod system saturation unless adaptation to the smaller background is prolonged. Based on previous studies, we suggest that the rod-desensitizing effect of small backgrounds under conditions of prolonged adaptation is due to a post-retinal mechanism.

Spatial and chromatic antagonism between the long- and middle-wavelength cones in the detection of long-wavelength flashes

Field additivity was tested for small red increments on small and large adapting fields. The results for a large adapting field showed the failure of field additivity of the cancellation type, providing strong evidence for cone-antagonistic coding, while the results for a small adapting field showed field additivity. The cancellation effect took place concurrently with the spatial sensitization effect, thus suggesting that the spatial and chromatic antagonism occur concomitantly at some stage in the detection pathway.

Increment spectral sensitivities for spatial periodic grating patterns: evidence for variable tuning of the chromatic system

Increment spectral sensitivities were measured for spatial periodic grating patterns. First, the increment threshold was determined as a function of wavelength, for various spatial frequencies and white-background intensities. Second, the additivity of test mixture was tested. Thirdly, the threshold vs intensity curves were determined for various spatial frequencies of test monochromatic stimuli. Finally, the increment threshold was determined as a function of spatial frequency. The following results were obtained: the background intensity and the spatial frequency affected the increment spectral sensitivity functions in different manners; the result of test mixture showed subadditivity for high background intensity and approximately linear additivity for low background intensity; the resultant t.v.i. curves converged toward Weber's law at high background intensity; and the sensitivity vs spatial frequency curve showed a loss of the sensitivity at low spatial frequency for high background intensity. These results called in question the existing hypothesis of separate chromatic and achromatic systems whose spectral tunings are invariant with a spatial parameter, but favored the hypothesis of variable tuning of the chromatic system.

Rod-cone interaction in flicker detection

There is considerable evidence in the literature that rod-cone interaction occurs when both rods and cones simultaneously detect a test target. More recent evidence, however, has shown a parafoveal rod-cone interaction during dark adaptation for a purely cone-detected flickering test stimulus; this influence on cone threshold appears to be mediated by surrounding rods. In this study, we demonstrate a similar rod-mediated influence on parafoveal cone-detected flicker threshold. More surprisingly, foveal cone-detected thresholds are also influenced by rods. This effect occurs over at least a 2 log unit intensity range of mesopic background level; cone-detected 25 Hz flicker sensitivity is enhanced by increasing the radiance of the background. The action spectrum of this effect fits the scotopic spectral sensitivity curve. At higher background levels, this rod-cone interaction disappears and surrounding cone activity then influences the cone flicker threshold. The results suggest that, as rods recover sensitivity, they reduce cone-detected flicker sensitivity, even at the fovea. The rod influence on cone flicker is most apparent for long wavelength test stimuli. Our results, in agreement with recent reports, suggest that the rod-cone interaction is laterally-mediated and may be specific for the long wavelength-sensitive cone type.

The cone threshold: spatial interactions of rod and cone adapting signals

Steady small blue adapting fields that stimulated only rods produced large threshold elevations for a superimposed tiny red flash that stimulated cones. The threshold elevation differed considerably between observers, confirming the results of Buck (Topical Meeting on Recent Advances in Vision, abstract in Technical Digest, 1980; Invest. Ophthal, visual Sci., Suppl. 20, 207, 1981). A red annulus that stimulated cones reduced slightly the threshold elevation produced by the small rod adapting field. Similarly, when the cone threshold was elevated by a small red field that stimulated only cones, a blue annulus that stimulated rods slightly reduced the threshold of the red cone test flash. These effects, although weak, demonstrate lateral sensitizing interactions between a cone center field and rod annulus, and vice versa, when the cone threshold is assessed with a tiny red flash.

Spatial sensitization and desensitization with small adapting fields: interactions of signals from different classes of cones

Thresholds for brief, tiny flashes were measured on small adapting fields at 5 degrees retinal eccentricity. Rod influences was eliminated by bleaching. Green and red small adapting fields raised the threshold for a green test flash by stimulating both middle and long wave cones-thus spatial desensitization by small adapting fields is not con-specific. However, a small adapting field that strongly stimulated only short wave cones did not affect the visibility of a long wave flash. Spatial sensitization was also measured, using green or red annuli that surrounded a small red field and green test probe. Sensitization was not cone-specific. Possible mechanisms producing desensitization with small fields and annular sensitization are discussed.