Saccadic latency effects of progressively deleting stimulus offsets and onsets.

We designed two extensions of Saslow's well-known gap and overlap conditions that require increased voluntary effort because of the progressive elimination of target onsets and fixation point offsets, and obtained repeatable data obeying simple numerical relations. For each of the five stimulus lighting conditions, saccadic latency was measured as a function of the retinal eccentricity or displacement of the target. Latencies were fitted by a serial processing model in which the visually guided minimum tracking latency VGL(min) is supplemented by two types of delay, dubbed 'unlock' and 'target', that can be either short or long ('direct' or 'indirect'), depending on the conditions. There are two findings: (1) The model has utility. The rank order of saccadic latencies for the five stimulus lighting conditions was constant across all subjects, sessions and eccentricities in the range 7.5'-6 degrees left or right. For pooled data, and the saccadic latency plateau (1-6 degrees), the model was also within +/-3 ms of the mean latencies. (2) Latencies of tiny saccades to intrafoveolar stimulation (7.5-45') were invariably long in all five stimulus conditions. One factor here must be the experimentally measured local prolongation of VGL(min).

The joint contributions of saccades and ocular drift to repeated ocular fixations.

We measured the joint contributions of different sized saccades and ocular drift to overall performance in an alternating fixation task. Subjects repeatedly shifted fixation between the centres of a pair of bars of width 2 degrees -2 arc min, either mentally selected or electronically cut from a stationary sine grating display. Eye movement patterns exhibited consistent features across all displays, and pairs of widely separated bars were studied most. Variability (S.D.) and relative accuracy (under/overshooting bias) were estimated from the concentration of eye positions over the two target bars. Overall variability, i.e. for eye movements as a whole, reached a minimum of 5 min for bar widths less than 20 min across subjects, displays and tasks. This was consistent, as were several other aspects of the study, with a constant 20-min diameter goal zone hypothesis. For wide bars, overall variability increased nearly as the square root of bar width, and for narrow bars, was independent of bar separation. A typical between-bar crossing saccade was tightly constrained in departure point but widely scattered in landing position, the associated variability increasing with bar separation. The final high overall precision was achieved largely by within-bar saccades of greater than 7.5 min effective amplitude that were present at a rate of 1 (range 0.3-3) per crossing saccade. This is consistent with views that very small saccades (the smaller microsaccades) make little obvious contribution to oculomotor performance.

Reliability and dimensionality of judgments of visually textured materials.

We extended perceptual studies of the Brodatz set of textured materials. In the experiments, texture perception for different texture sets, viewing distances, or lighting intensities was examined. Subjects compared one pair of textures at a time. The main task was to rapidly rate all of the texture pairs on a number scale for their overall dissimilarities first and then for their dissimilarities according to six specified attributes (e.g., texture contrast). The implied dimensionality of perceptual texture space was usually at least four, rather than three. All six attributes proved to be useful predictors of overall dissimilarity, especially coarseness and regularity. The novel attribute texture lightness, an assessment of mean surface reflectance, was important when viewing conditions were wide-ranging. We were impressed by the general validity of texture judgments across subject, texture set, and comfortable viewing distances or lighting intensities. The attributes are nonorthogonal directions in four-dimensional perceptual space and are probably not narrow linear axes. In a supplementary experiment, we studied a completely different task: identifying textures from a distance. The dimensionality for this more refined task is similar to that for rating judgments, so our findings may have general application.

Fixation conditions, the foveola and saccadic latency.

Saccades are often elicited in the laboratory by the abrupt step-displacement of a single lit point which is initially the foveolar fixation point and then the eccentric refixation target. This was our Control condition. Four experiments modified the fixation arrangements to examine the effect of altered foveolar stimulation on saccadic latency and accuracy to targets within the central +/- 6 deg of the visual field. (1) No foveolar fixation point: The subject fixated the empty space midway between a pair of fixation guides, which later collapsed into a single refixation target. Latencies for small saccades were similar to the Control values. (2) No foveolar fixation point and no real refixation target: A pair of fixation guides underwent a yoked displacement, and it was easy to fixate and track the invisible midpoint. The smallest saccades were hypermetric, and the typical pattern of latency variation with retinal eccentricity was exaggerated in scale. (3) Spatial effects of a persistent non-target: The precise position of a non-target was important, latency increases being in the ipsilateral hemifield when the non-target was intrafoveolar and unilateral, bilateral when intrafoveolar and on the midline, and local when the non-target was extrafoveolar. (4) Temporal effects of a foveolar fixation point: Blanking an otherwise persistent fixation point for as little as 1 msec at the time of target presentation reduced the expected latency increase. We conclude that the position and timing of foveolar illumination can be critical for saccades of all sizes.

Saccades to large coloured targets stepping in open fields.

Subjects tracked a 2.3 deg target that stepped 5 deg, in a randomly chosen direction, each time it was foveated. Targets were coloured patches that were fairly close to white; in some cases precise matches ensured equiluminosity with the background. Viewing conditions provided good colour rendering and neutral colour adaptation. Pale colours were surprisingly well tracked. Multiple regressions showed that the colour and spatial characteristics of the target are important determinants of a primary saccade's latency. Significant factors included target size, achromatic contrast, tritanopic purity difference, and chromatic saturation. Colour-normal subjects always responded more slowly to yellow or blue targets which a deuteranomalous subject tracked quite well. Severely blurring the target had a consistent minor effect.

Retinal eccentricity and the latency of eye saccades.

The latency-eccentricity function, for the first saccade towards a small green or red target on the horizontal hemi-retinal meridian, is bowl-shaped with a central latency peak, a minimum plateau from 0.75 to 12 degrees, and a gradual increase in latency towards the periphery. The function is highly reproducible and the central peak is a robust finding. For our step sizes and criteria its height is around 35-75 msec. Manipulations of target intensity and colour, or the state of adaptation by a retinal bleach, show that sensory contributions to the central peak are generally small (5-15 msec) for adequately suprathreshold targets. Beyond 35 degrees in the temporal retina (nasal visual field) latencies become erratic, even for bright green targets, and there are direction errors. Frequent direction errors occur over a much wider range of eccentricities for red targets that are 1 log above the foveal threshold for perception.

Preattentive discrimination of relative phase modelled by interacting Gabor or by difference-of-Gaussian filters.

A computer simulation of the human preattentive visual pathway used Gabor and difference-of-Gaussian (DOG) filters to model two-dimensional relative phase discrimination. There is a hierarchy of hexagonally packed levels: (i) an image layer; (ii) an intermediate level of on- and off-centre DOG filters; (iii) a partial set of broadband oriented Gabor-like filters with high-level DOG filters in parallel. Connections between layers use half-wave rectification and a compressive nonlinearity. Local feedback interactions between oriented Gabor filters, together with spatial averaging, allow the model to discriminate both two-dimensional relative phase and orientation differences. There were two separate simulations for the Gabor filters, one with even-symmetric filters and another with odd-symmetric. Textures were formed by superposing three high contrast sine-wave gratings with successive rotations of 60 degrees. Relative phase and global orientation were the varied parameters. Psychophysical rating data for peripheral viewing of texture pairs resemble the results from the even-symmetric simulation, showing discrimination of relative phase and orientation. In contrast, the DOG filters in the model simulate only the relative phase aspects of the data.

Texture segregation based on two-dimensional relative phase differences in composite sine-wave grating patterns.

Experiments examined the visual processing of relative phase relations between differently oriented components of a textured pattern. The textures were a super position of three 1.3 c/deg sine-wave gratings rotated 60 deg relative to one another. Global orientation and relative phase were varied. Subjects rated the segregation of pairs of textures presented in a figure/ground configuration at 10 deg retinal eccentricity. Both orientation and phase differences were used in making ratings. It was not simply the difference in relative phase that mattered. The results also depended on the values of the relative two-dimensional (2-D) phases in figure and ground. Mirror-image texture pairs (which may have a large relative phase difference) segregated poorly compared to other pairs with similar phase differences. This suggests that the peripheral visual system does not completely encode 2-D relative phase. Secondary issues include spatial frequency effects, the relevance of figure/ground borders, and better performance in the lower visual field.

Segregation of mesh-derived textures evaluated by resistance to added disorder.

In the "figure detection task" the strength of segregation for a particular texture pair was estimated by the threshold amount of added disorder that prevented segregation of a textured figure from a textured ground. Disorder was either jitter in the orientation of the texture elements, or jitter in their xy positions, or a mixture of the two. Other procedures included lowpass filtering, and a task requiring discrimination between textured figures of different shapes. Orientation cues are weakly or inconsistently used for segregating mesh textures. The low spatial harmonics are very important. A new finding is that orientation and position jitter thresholds for a set of figure/ground texture patterns are often proportional. In a mixture the one disorder can be exchanged for the other.

The discrimination of blur in peripheral coloured borders.

Our interest is in measuring the threshold for blur, for long borders at well-defined retinal eccentricities between large fields of colour. Our method is to precisely match a centrally fixated sharp-edged disk of one colour against a surround of another, using the minimum distinct border criterion. We then measure the amount of added Gaussian blur that makes the sharp step-edged border just noticeably different. The selected colours are small fixed chromatic deviations of 500 msec duration from an adaptive low photopic or mesopic white. The effects of retinal eccentricity are striking. Blur thresholds are large for yellow-blue borders centrally (at 1.25 degrees eccentricity) and extremely large for red-green borders peripherally (at 20 degrees). Blur thresholds are generally larger for isoluminous chromatic borders than for a low contrast achromatic border. Accurate Rovamo-Virsu M-scaling is limited to photopic achromatic borders.

Segregation of some mesh-derived textures evaluated by free viewing.

Beginning with a hexagonal mesh, a family of equiluminous geometric textures can be formed by rotating each line segment by a fixed amount. Such textures vary both in line orientation and in the sizes of the intervening spaces or "holes". In the Fourier domain the textures have the same discrete two-dimensional frequency components but the amplitudes vary. The segregation of texture pairs as figures in grounds has been assessed for free viewing by the ratings of 10 subjects. Line segment orientation information alone cannot explain the results. Linear regression techniques were used to examine the correlation between the subjects' ratings and physical differences between figures and grounds. Significant correlations were found with (a) the amplitudes of the low-frequency Fourier harmonics of the textures and (b) maximum hole size, but not with the third harmonics or line segment orientation. Patterns composed of mirror-image pairs are very poorly discriminated; in these cases line orientation information may be used.

Detection of briefly flashed sine-gratings in dark-adapted vision.

Scotopic contrast sensitivity was measured near 20 deg retinal eccentricity for briefly flashed (10 or 20 msec) sine-wave gratings presented in darkness to dark-adapted subjects. For very low spatial frequencies (0.2-0.5 c/deg), curves of contrast sensitivity vs luminous energy show evidence of a low rod plateau and a high scotopic region, with an intervening transition at around -2 to -2.5 log scot td sec. Similar measurements made using long flashed or flickering gratings do not show a plateau. The results suggest that vision in the low rod region is impaired for brief flashes. For the briefly flashed stimuli, curves of contrast sensitivity versus spatial frequency in the low region were best fit by simple Gaussian functions with a variable centre size (sigma c = 0.5----0.25 deg), size decreasing with increasing flash energy. Difference-of-Gaussian functions with constant centre size (sigma c = 0.25 deg) provided the best fit in the high region. Overt input from the cones and grating area artefacts are excluded by appropriate tests. Calculation of photon flux into the receptive field centres suggests that signal compression in P alpha ganglion cells contributes to the low rod plateau.

Variable contributions of rods and cones to saccadic eye-movement latency in a non-foveating task.

Latencies were measured for anti saccades away from a small lit cue that steps +/- 10 deg in complete darkness. Cue luminance and wavelength were varied. Additional measurements were made during dark-adaptation or on backgrounds or at different retinal eccentricities. Luminance matched cues and Palmer's equivalent luminance transformation were also used. Latencies for pure rod and pure cone inputs obey Piéron's law in much the same manner as foveating saccades, except that latencies are somewhat longer. However, as judged by saccadic latency, interaction between rods and cones is quite variable in the anti task. The rod-cone transition either occurs at cue luminances well above the cone threshold and is from pure rod input to primarily cone, or occurs at the cone threshold and is from rods to rods-plus-cones. Direction errors, or reflex foveating saccades, are particularly increased for mesopic cues. The variable behaviour of subjects at the anti task is discussed in relation to temporal multiplexing of rod and cone signals from dark-adapted retinal ganglion cells, the delaying nature of the task, and attentional mechanisms.

Dependence of saccadic eye-movements on stimulus luminance, and an effect of task.

Various aspects of saccadic eye-movements are related to stimulus luminance for a small lit stimulus that steps 10 degrees horizontally in complete darkness. The relations depend on whether the stimulus is the target for a foveating saccade, or is the cue for an "anti" saccade which peripheralizes the retinal image of the cue: (1) at scotopic luminances the differences between foveating and anti saccades are diminished, largely because foveation is the more severely affected. Latencies are long amplitudes are scattered, and direction errors are not infrequent in both tasks; (2) the latency-luminance relation for foveating saccades shows an abrupt discontinuity at the perceptual rod-cone transition. Above the cone threshold corrective secondary saccades appear in greater numbers; (3) the corresponding latency transition for anti saccades is anomalous and protracted. Latency remains constant for mesopic cue luminances up to 1.0 log unit brighter than the perceptual rod-cone threshold. Direction errors are especially common in this mesopic luminance range. Mechanisms are discussed.

Rod-cone dependence of saccadic eye-movement latency in a foveating task.

This study examines the relations between some well known oculomotor functions (saccades) and well known retinal physiology (dark adaptation): it deals with the overall latency versus target luminance functions, with the underlying rod and cone latency-luminance functions, and with the synergistic interaction between these latency functions for mesopic targets. Saccadic latency was measured to small lit targets presented at 10 deg retinal eccentricity in complete darkness. Target luminance and wavelength were varied. Additional measurements were made during dark adaptation or on backgrounds, or at different retinal eccentricities. Luminance matched stimuli and Palmer's (1968) equivalent luminance transformation were also used. Latency is determined by an achromatic luminance mechanism that receives substantial rod inputs above the cone threshold. Latencies for pure rod or pure cone inputs increase rapidly as target luminance decreases. For the rods this latency increase appears to represent the waiting time for the 140 or so photons (lambda = 507 nm) that are required for a saccade. Errors in direction occur at scotopic luminances, or at low photopic luminances when only cones are functioning.

Quantum efficiency of dark-adapted human vision.

In the classic frequency-of-seeing experiment with a small brief stimulus, the overall efficiency F is 0.06 (lambda = 507 nm) in conventional long experiments and 0.09 in pooled short experiments. These results can be reconciled by postulating higher-level effects in the long experiments, which increase variability without substantially changing mean threshold. The quantum efficiency of dark-adapted vision Fd, the fraction of corneal photons acting at the decision level, is estimated to be 0.13-0.20 (maximum range, 0.12-0.30).

The differentiation of visually guided and anticipatory saccades in gap and overlap paradigms.

Saslow and others have shown that the latency of foveating saccades can be altered by changing the offset time of the current fixation point relative to the onset of the peripheral target. Whether anticipatory saccades contributed to these results was not known. By the criteria of direction error and amplitude error the minimum latency for visually guided saccades is 110-130 ms for three subjects and 160 ms for a longer latency subject. Excluding anticipatory responses did not eliminate offset-onset effects. The genesis of express saccades and the role of higher neural levels is discussed.

The scale of the visual pathways of mouse and rat.

Photoreceptors and neurons at various levels to cortex have been counted in mouse and rat. The ratios of neuron numbers (rat/mouse) are similar to the ratio of retinal areas or the squared ratio of eye sizes; so to a first approximation the two species have linearly scaled eyes, equal photoreceptor spacings (in micron), and visual pathways scaled numerically by the number of photoreceptors. With supplementary data from the literature, some of the functional implications of the design can be evaluated level by level. Overall, there is structural and computational economy, or even parsimony.

On plotting amplitude-transition functions for voluntary eye saccades.

The first of two oppositely directed saccades may be small if the time interval between the two saccades is short. We have examined the merits of plotting the amplitude of the first saccade against saccadic interval. This plot provides a useful supplement to other approaches in the literature, as it is insensitive to variation due to extraneous factors or to task. Reasons for this are discussed.

Power spectra for ocular drift and tremor.

Eye position during fixation has been measured without contact and the power spectrum of drift and tremor eye movements has been estimated. In the 0-40 Hz frequency range power declines with frequency roughly as 1/f2. In the 40-100 Hz frequency range the dominant spectral component is a broad spectral peak with a peak amplitude of about 6 arc sec. The shape of the broad spectral peak is dependent on the fixation direction. A model that explains qualitatively the shape of the power spectra function is presented. It is suggested that tremor eye movements are by-products of the clock-like firing of motor neurons.