Vision in the presence of known natural retinal image motion.

Previously we reported that failures of compensatory eye movements led to appreciable binocular retinal image motion during head rotation. Subjectively, the visual world appeared clear, fused, and stable under these conditions. The present experiments examined these impressions psychophysically. The spatial modulation transfer function of subjects with known retinal image motion was measured during head rotation. We found that contrast sensitivity was reduced for gratings over 6 cycles/degree and was increased for lower spatial frequencies. Our results, when compared with Kelly's [J. Opt. Soc. Am. 69, 1340-1349 (1979)] measurements made with artificially moving stabilized gratings, show that natural retinal image motion is less harmful to contrast sensitivity at high spatial frequencies and more beneficial at low spatial frequencies. Furthermore, we had previously found that natural retinal image motion was different in each eye during head movement but no diplopia was noticed. We confirmed this subjective impression by measuring forced-choice stereoacuity thresholds concurrent with binocular head and eye recordings. Stereoacuity was not disturbed by large fixation disparities or high vergence velocities. Recordings also were made while a fused Julesz stereogram was viewed during attempts to break fusion with violent head movements. Fusion could not be broken. Stereograms turned on during violent head movement fused rapidly. We conclude that vision is better with natural retinal image motion than expected from experiments done with stabilized heads.

Receptive-field symmetry probed using converging gratings.

Contrast-sensitivity functions (CSF's) for converging and diverging gratings were obtained under voluntary fixation at several field widths and under retinal stabilization at one field width. In general, these types of gratings had similar CSF's at all temporal frequencies tested, a result that can be explained by plausible spatiotemporal receptive-field models possessing spatial antisymmetry but not by models possessing pure (even) spatial symmetry. Phase condition at the central line in converging and diverging gratings affected grating detectability at high spatial frequencies, as expected.

Phase shift in red and green counterphase flicker at high frequencies.

When balanced red and green lights are alternated more than 20 times per second, the perceived flicker can be reduced by advancing the green flicker about 10 degrees of the red-green cycle. The required advance for least flicker is greatest at retinal illuminances around 1000 td and frequencies between 30 and 35 Hz. A model that predicts tuning at this frequency exists, but the tuning curve that is predicted is broader than that observed. A modified model is left for future publication. Meanwhile, other empirical properties of the advance required by green over red are described. In addition to the intensity dependence of this phase shift, we describe its dependence on intensity balance between red and green. Also, the intensity balance turns out to depend on the frequency being used, in contrast to the independence expected by Ives, the inventor of heterochromatic flicker photometry.

Limulus brain modulates the structure and function of the lateral eyes.

At night efferent optic nerve activity generated by a circadian clock in the Limulus brain changes the structure of the photoreceptor and surrounding pigment cells in the animal's lateral eyes. The structural changes allow each ommatidium to gather light from a wider area at night than during the day. Visual sensitivity is thereby increased, but spatial resolution is diminished. At daybreak efferent activity from the clock stops, the structural changes reverse, and the field of view of each ommatidium decreases. The cyclic changes are endogenous and continue in the dark. Thus, under the control of a circadian clock, the Limulus eye exchanges its daytime acuity for greater sensitivity at night.

Shifts in perception of size after adaptation to gratings.

After viewing a suitable grating of vertical stripes for 5 minutes, subjects overestimated the width of a rectangle by 6 percent. The shifts in perception of size occurred whether individual stripes in the grating were narrower than, equal to, or wider than the rectangle. Rectangle width was underestimated only if the grating stripes were extremely wide, with a spatial frequency lower than most of the effective amplitude spectrum of the rectangle. These findings (and complementary ones with horizontal gratings) suggest that the visual system codes size on the basis of spatial frequency components, rather than directly in terms of width.

Perception of size of one object among many.

Adaptation to a grating of properly chosen frequency may lead to two apparently conflicting observations: Another grating may then appear to be of increased frequency (compared with its "unadapted" frequency) while the individual bars of the grating appear to have widened. This perceived widening parallels previous results with single bars. By attending to only one grating bar, the subject effectively seems to change the grating frequency spectrum to that of a single bar.

Visibility of aperiodic patterns compared with that of sinusoidal gratings.

1. Using experimental curves relating the threshold contrast of sinusoidal grating patterns to their spatial frequency, the expected threshold contrast curves for three aperiodic patterns, viz. a single half-cycle sinusoid bar, a single full-cycle sinusoid bar, and the boundary between an extended sinusoidal grating and a 50% grey surround, are calculated. In this calculation the assumption is made that the system is linear near the threshold.2. Experiments are described in which the actual threshold contrast curves are determined for these aperiodic patterns by three observers. The patterns were generated on the face of an oscilloscope and could be varied in size and contrast.3. These experimental curves agree well with the predictions in the high frequency region (i.e. above about 10 c/deg), but below this various complicating factors restrict the validity of the calculations.4. Thus there is no reason to suppose that a linear theory cannot be used to predict visibility of aperiodic patterns near threshold.

Flicker fusion phenomena.

The high-frequency temporal behavior of the human visual system has been shown to have some of the properties of a linear low-pass filter. For such a system it is appropriate to consider a repetitive stimulus as having separable Fourier harmonic components. The direct-current component or average luminance is important in that it sets the adaptation level. It is therefore convenient to keep it constant when varying other stimulus parameters, such as frequency or wave form. Of the alternating-current components, only the fundamental is important at high frequencies, the higher harmonics being relatively more attenuated. Any linear low-pass filter system responds in a predictable way to sinusoidal stimulation, whether continuous or of short duration. In the case of the visual system, predictable behavior is found at high frequencies, and it leads to discovery of hitherto unobserved pseudoflash and real flash phenomena. Measurement of a new characteristic time is suggested. At low frequencies the use of half and whole-sinusoidal flashes leads to the discovery of some interesting relations with flicker thresholds, but these remain for future discussion.