Gestalt factors modulate basic spatial vision.

Human perception of a stimulus varies depending on the context in which the stimulus is presented. Such contextual modulation has often been explained by two basic neural mechanisms: lateral inhibition and spatial pooling. In the present study, we presented observers with a vernier stimulus flanked by single lines; observers' ability to discriminate the offset direction of the vernier stimulus deteriorated in accordance with both explanations. However, when the flanking lines were part of a geometric shape (i.e., a good Gestalt), this deterioration strongly diminished. These findings cannot be explained by lateral inhibition or spatial pooling. It seems that Gestalt factors play an important role in contextual modulation. We propose that contextual modulation can be used as a quantitative measure to investigate the rules governing the grouping of elements into meaningful wholes.

The effect of spacing regularity on visual crowding.

Crowding limits peripheral visual discrimination and recognition: a target easily identified in isolation becomes impossible to recognize when surrounded by other stimuli, often called flankers. Most accounts of crowding predict less crowding when the target-flanker distance increases. On the other hand, the importance of perceptual organization and target-flanker coherence in crowding has recently received more attention. We investigated the effect of target-flanker spacing on crowding in multi-element stimulus arrays. We show that increasing the average distance between the target and the flankers does not always decrease the amount of crowding but can even sometimes increase it. We suggest that the regularity of inter-element spacing plays an important role in determining the strength of crowding: regular spacing leads to the perception of a single, coherent, texture-like stimulus, making judgments about the individual elements difficult.

Attention and perceptual learning modulate contextual influences on visual perception.

Brightness discrimination thresholds and facilitation by lateral interaction were measured in five human observers and two monkeys. The subjects judged the brightness of one of four peripherally seen lines against a reference. This experiment was performed both when the observer was cued to the position of the test line (focused attention) and when there was no cue (distributed attention). Discrimination was better with focused than with distributed attention. When the test line had a collinear flank, its brightness was enhanced; this enhancement was four times more prominent with distributed than with focused attention. After training, thresholds improved and collinear facilitation decreased under distributed but not under focused attention. The findings show that there are fewer benefits from contextual interaction once attention is directed toward a visual location, and that the attentional effects are subject to training.

Improvement in visual sensitivity by changes in local context: parallel studies in human observers and in V1 of alert monkeys.

To explore the role of primary visual cortex in contour integration, we measured the contextual sensitivity of human contrast thresholds and of superficial layer complex cells in monkey V1. An observer's contrast detection was 40% improved by a second suprathreshold bar; the effect was decreased as the two bars were separated along their axis of orientation, were displaced from colinearity, and had their relative orientation changed. Recordings from V1 showed that 42% of complex cells demonstrated facilitation for a second bar outside their classical receptive fields with a similar dependency on relative location and orientation. Both effects were eliminated by an orthogonal line between the two iso-oriented lines. Multiple randomly placed and oriented lines in the receptive field surround often caused a reduction in a cell's response to an optimally oriented stimulus, but this inhibition could be eliminated by changing the orientation of a few of these elements to colinearity with the centrally located target.

Vertical disparity detection: is there an induced size effect?

Horizontal, vertical, and over-all size differences were introduced in the retinal images of the two eyes of normal subjects during brief presentations of simple foveal targets. Horizontal disparities, whether accompanied by vertical disparities or not, induced the appearance of a rotation of the target around a vertical axis out of the frontal plane, according to the expectation from geometry, but vertical disparities had no effect. Over-all size changes in one eye induced the effect of the horizontal component. Threshold experiments showed that even with practice and error feedback, vertical disparity detection has at most only one-tenth the sensitivity of horizontal disparity detection. Although at variance with findings on the induced size effect obtained under more complex observation conditions, these results confirm that the processing of horizontal disparity plays a special role in the integration of the signals coming from the two eyes.

Evaluating diffusion of light in the eye by objective means.

PURPOSE: The authors have developed an index of diffusion that describes the relative spread of light inside and outside the region of image focus of the living human eye. It provides in numerical terms a measure of light scatter and can be used to characterize the optical deficit in eyes with age- and disease-related abnormalities of the anterior segment. METHOD: An improved version of the double-pass method of examining the aerial image formed by reflection of the retinal image of a point source is employed, together with a new way of analyzing the image. Experimental estimation shows the contaminating effect of back scatter from the media and corneal reflection to be negligible. Measurements are objective and do not require any responses on the part of the patient. Data become available practically on-line. RESULTS: Index of diffusion values were obtained on 13 patients and varied from 0.22 to 1.04, strongly tending to increase with age. They are rather robust to pupil size, exposure duration, and small amounts of defocus. CONCLUSION: The index appears to provide a promising measure of optical performance of the media of the anterior segment of the eye, which might be useful in studying the effect of aging, injury, and disease.

Stereogram design for testing local stereopsis.

The basic ability to utilize disparity as a depth cue, local stereopsis, does not suffice to respond to random dot stereograms. They aslo demand complex perceptual processing to resolve the depth ambiguities which help mask their pattern from monocular recognition. This requirement may cause respnse failure in otherwise good stereo subjects and seems to account for the long durations of exposure required for target disparities which are still much larger than the best stereo thresholds. We have evaluated spatial stimuli that test local stereopsis and have found that targets need to be seth optimal separation of components and brief exposure duration (250 msec), thresholds fall in the range of 5 to 15 sec arc. Intertrial randomization of target lateral placement can effectively eliminate monocular cues and yet allow excellent stereoacuity. Stereoscopic acuity is less seriously affected by a small amount of optical defocus when target elements are widely separated than when they are crowded, though in either case the performance decrement is higher than for ordinary visual acuity.

What prior uniocular processing is necessary for stereopsis?

In order to induce stereoscopic depth, retinal images in the right and left eyes must differ. Our experiments demonstrate that these uniocular differences for simple three-line configurations at stereo-threshold cannot usually be distinguished one from the other. We confirm that depth values are associated with individual features by virtue of their disparity, rather than by a comparison of their image separations in the two eyes. In ordinary stereoacuity patterns, no prior identification of monocular relationships is needed to produce discriminable depth differences. For certain patterns, disparity of the positions of the averages of the targets' internal light distributions in the two eyes can substitute for disparity of pattern contours in stereoscopic depth discrimination.

The Ferrier Lecture, 1992. Seeing depth with two eyes: stereopsis.

The primate's two eyes both point forward, and are a few centimetres apart. The task of extracting depth from two slightly dissimilar two-dimensional retinal images falls on the central nervous system. The mechanism is so effective that a good observer can detect, by stereopsis alone, depth differences of as small as 10 microns at a distance of 25 cm. Psychophysical and neurophysiological studies have begun to unravel the neural circuits in the brain which process the retinal signals to yield stereoscopic information.

Location and line orientation as distinguishable primitives in spatial vision.

In geometry the same parameters define a straight line segment as its two endpoints, yet the human observer's perception of spatial relations can differ depending whether an actual line is shown or merely its terminators. Experiments are reported that demonstrate that the position and orientation senses can be largely decoupled. Because both point localization and line orientation each in its own way has a quality of immediacy, yet manifests high performance in the hyperacuity range, both must be regarded as a primitive each with its own neural processing mechanisms. This can give rise to perceptual dissonances (visual illusions) of which one example is analysed.

Visual discrimination of fractal borders.

The ideas of fractals and fractal dimension are here translated into the realm of visual psychophysics. Borders between two fields of different luminance were used. Because of the finite grain of the visual system, fractal dimension need be defined only within a certain size range. For a fractal dimension of 1.15, the just-detectable difference in fractal dimension was found to be about 0.0085, rising to about 0.015 for a fractal dimension of 1.25. Reducing exposure duration from 1 s to 0.33 s decreases sensitivity to differences in fractal dimension, but there was no gain in increasing the exposure duration. Good visual observers who are naive to the task require some training before reaching optimal performance. The ability to discriminate fractal dimension differs between fractal edges of the same fractal dimension that were generated with differing statistical programs. Even after considerable training, an observer makes 29% errors when asked to distinguish a fractal edge generated with a Gaussian random walk from one with a rectangular random walk. Gaussian random walk fractals can be more easily distinguished from Poissonian and Cauchy ones.

Relative localization in the human fovea: radial-tangential anisotropy.

A distinction is drawn between two intrinsic directions within a simple spatial configuration. The line joining two elements is the radial direction and orthogonal to it is the tangential direction. Separation and bisection discrimination and just-detectable differences in line length are examples of radial thresholds. Vernier and alignment detection are tangential thresholds. Neural processing along these two intrinsic directions differs. There is a strong 'oblique effect' for tangential thresholds and virtually none for radial thresholds. Flank interaction impairs tangential but not radial thresholds.

Detection and processing of vertical disparity by the human observer.

Based on the distinction between uniocular vertical magnification and vertical disparity, the induced size effect experiments were reinterpreted and new experiments done to show that vertical disparity signals can produce other stereoptic depth effects. The direction and efficiency of utilization of vertical disparity signals depend on the quadrant of the visual field and the stimulus position within it.

Contrast and duration of exposure differentially affect vernier and stereoscopic acuity.

Although stereoacuity and vernier acuity both yield comparable thresholds well below the eye's resolution limit, the neural circuits for these two classes of visual responses do not process the signals in an identical manner. It had previously been demonstrated that hyperacuity is more resistant to image blur than stereoacuity and that the zones within which two targets must be placed to achieve the lowest thresholds differ quite radically. Two further differences are reported here: reduced contrast affects stereoacuity more severely than hyperacuity, as also does shortening of exposure into a range of tens of milliseconds, even when the Bunsen-Roscoe-Bloch law has been factored out.

Is peripheral visual acuity susceptible to perceptual learning in the adult?

While it is generally accepted that foveal visual acuity in the adult has reached an optimal value, claims for improvement of peripheral acuity with training in the adult persist in the literature. Practice effects in peripheral hyperacuity have been amply documented. A carefully controlled test is here reported to examine the influence of training on the resolution thresholds for two lines and on Landolt C acuity measurements in the retinal periphery in eight normal adults. It involved 11-30 daily sessions of 300 responses with feedback. In some observers the first day's results were somewhat poorer, but otherwise the threshold curves were essential flat. Yet in the same location vernier acuity could be improved by 50% in six training sessions. Sustained and lasting neural modifications in peripheral vision can take place in stereoscopic, orientation, vernier, bisection and time discriminations, but not in resolution and Landolt C acuities.

Lines and Gabor functions compared as spatial visual stimuli.

Foveal discrimination thresholds were measured for orientation, vernier alignment, bisection, displacement detection and stereoscopic acuity using simple line stimuli and also sinusoidal grating patches with gaussian intensity modulation (Gabor stimuli). Stimulus parameters such as luminance, exposure duration, component separation and, as far as possible, length were identical for both. Orientation discrimination as a function of length is almost identical for the two classes of stimuli, some observers performing slightly better with Gabor patches. Thresholds for displacement detection are also the same. Vernier, stereo and bisection acuities, however, are considerable better with line than with Gabor stimuli.

The oscilloscopic view: retinal illuminance and contrast of point and line targets.

The specification of intensity of self-luminous point and line targets, such as those created on cathode-ray tubes, is discussed and methods are outlined for measuring it and deducing the resulting retinal illuminance. The calculations can be extended to the specification of contrast by combining them with the more traditional method of identifying the retinal illuminance provided by uniform backgrounds. Such values of contrast are not independent of observation distance.

The spatial grain of the perifoveal visual field.

Measurements for normal human photopic vision in the fovea and in 2.5, 5 and 10 degree positions in the periphery reveal that hyperacuity thresholds rise faster with eccentricity than visual resolution thresholds. On the other hand, the retinal distances over which hyperacuity detection remains optimal show a much slower rise. At 10 degrees eccentricity, hyperacuity thresholds are 10 times higher visual resolution thresholds 4-5 times higher, and optimal processing distances for hyperacuity only 2-3 times larger than at the fovea. Psychophysical procedures do not, therefore, provide a single unambiguous measure for the changes of spatial grain across the visual field.