Speed, spatial-frequency, and temporal-frequency comparisons in luminance and colour gratings

The perceived speed, temporal frequency, and spatial frequency of matched colour and luminance gratings were compared in separate experiments. The large factor by which colour gratings are perceived to be slower moving than matched luminance gratings cannot be explained by systematic differences in the perceived spatial frequency or in the perceived temporal frequency of the two types of grating.     

Reaction time to motion onset of luminance and chromatic gratings is determined by perceived speed

We measured reaction times for detecting motion onset for sinusoidal gratings whose contrast was modulated in either luminance or chromaticity, for various drift rates and contrasts. In general, reaction times to chromatic gratings were slower than to luminance gratings of matched cone contrast, but the difference in response depended critically on both contrast and speed. At high image speeds there was virtually no difference, whereas at low speeds, the difference was pronounced, especially at low contrasts. At high image speeds there was little dependence of reaction times on contrast (for either luminance or colour), whereas at low speeds the dependence was greater, particularly for chromatic stimuli. This pattern of results is reminiscent of those found for apparent speed of drifting luminance and chromatic gratings. We verified the effects of contrast on perceived speed, and went on to show that the effects of contrast on reaction times are totally predictable by the perceived speed of the stimuli, as if it were perceived rather than physical speed that determined reaction times. Our results support that idea of separate systems for fast and slow motion (with separate channels for luminance and colour at slower speeds), and further suggest that apparent speed and reaction times may be determined at a similar stage of motion analysis.     

Thresholds for the identification of the direction of motion of plaid patterns defined by luminance or chromatic contrast

Contrast thresholds for identification of the direction of motion were determined for sinusoidal gratings and plaid patterns moving in eight possible directions. Since plaid patterns are the sum of two component gratings, a prediction of the thresholds for plaids can be made by assuming that the motions of both component gratings are independently identified (probability summation). In agreement with standard two-stage models of plaid perception, our results show that for stimuli defined by luminance contrast, plaid direction thresholds can be predicted well from the component thresholds. This also holds for fast-moving isoluminant plaid patterns, but for slowly moving (< 4 Hz) isoluminant plaids, direction thresholds were substantially higher than the prediction from the components. In the latter case, subjects frequently were unable to identify the motion of the plaid in the pattern direction, even when the direction of motion of both components could be reliably identified. Different mechanisms might underlie the perception of luminance and isoluminant plaids at slow speeds.     

Estimating the gradient direction of a luminance ramp

We report on the extent to which human observers are able to indicate the gradient direction of a luminance ramp. In our experiment modulation depth ranged from 1 to 64% and field sizes subtended 0.5 to 47.5 degrees of visual angle. Observers are not able to indicate the gradient direction for modulation depths below 6%. For values above this threshold, the angular standard deviation in the responses decreases proportionally with the logarithm of the modulation depth and is about 11 degrees for a modulation depth of 64%. The angular standard deviations for supra-threshold contrast are slightly increased for the field sizes of 0.5 and 47.5 degrees but are constant over a range of field sizes of 1 to 25 degrees. Thus, size invariance holds well for this range of field sizes.     

The interaction between stereoscopic and luminance motion

An interaction in apparent motion between perceived three-dimensional forms defined by stereopsis and local luminous elements is reported. Vertical stripes of cyclopean square gratings were simulated by random-dot stereograms. Alternation of two-frame stereograms whose phases differed by 90 deg caused two kinds of percepts, planes' motion in depth (first-order stereoscopic motion, first-order SM) or lateral motion of gratings (higher-order stereoscopic motion, higher-order SM). Experiment 1 explored the conditions under which higher-order SM frequently arose, as opposed to local luminance-based in-depth motion (first-order SM). The results show that, when the spatial arrangements of two-frame random dots were correlated, higher-order SM dominated for long ISI conditions (ISI > 73 msec). When they were uncorrelated, higher-order SM dominated even under zero ISI conditions. Subjects reported that, when higher-order SM was seen, dots were attached to the surfaces of the moving cyclopean figure (motion capture). Experiment 2 tested which factor caused the domination of higher-order SM under uncorrelated conditions in Experiment 1, the larger distance of dot jump or the varied directions of the dots' motion. The results show that, when the distance of dot jump is large or when the directions of dots' motion are incoherent, higher-order SM arises more frequently. When local first-order motion signals are weakened by appropriate temporal and spatial conditions or by incoherent motion directions, higher-order SM dominates and it captures the motion of dots.     

Representations of motion and direction.

In 6 experiments, incidental memory was tested for direction of motion in an old–new recognition paradigm. Ability to recognize previously shown directions depended greatly on motion type. Memory for translation and expansion–contraction direction was highly veridical, whereas memory for rotation direction was conspicuously absent. Similar results were obtained in conditions in which motions were illustrated with pictures. Results suggest that explicit representations of direction in long-term memory are not so much related to motion per se as to the consequences of motion, the displacements of objects. Memory for all motions following circular pathways was found to be corrupted by a generic bias to regard the clockwise direction as familiar. Assessment of memory in these cases required disentangling familiarity bias for the clockwise direction from explicit recognition of direction. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Motion capture of luminance stimuli by equiluminous color gratings and by attentive tracking

Two experiments demonstrated motion capture of luminance-defined dots by gratings with no net luminance-based motion. In a series of two-frame experimental trials, we superimposed bright dots and a color grating rotating in opposite directions. Capture was observed at equiluminance and was facilitated by the presence of color in gratings over a range of luminance contrasts. In a second experiment, observers noted that when a counterphase grating was tracked in either direction with attention, the superimposed dots were captured in that direction. These results suggest motion capture is supported not only by luminance-based motion, but also by color- and attention-based motion. Indeed, we suggest that the most parsimonious explanation is that all capture is mediated by attention.     

Directions of motion after-effects induced by gratings and plaids

In three experiments the direction of motion after-effect (MAE) is measured following adaptation to two gratings moving in different directions presented in alternation (component-induced MAEs: CMAEs), and to moving plaid patterns composed of superimposed pairs of these gratings (plaid-induced MAEs; PMAEs). These MAEs are compared to: (i) the vector sum direction of the component gratings; (ii) the IOC-predicted direction of the plaids; and (iii) the perceived direction of the plaids as reported by observers. Contrary to previous findings (Burke D, Wenderoth P. Vis Res 1993;33:351-9), directions of PMAEs are shown to approximate the vector sum direction of the components, whereas directions of CMAEs are shown to approximate the mean (unweighted) direction of the components. This difference is attributed to the activity, and adaptation, of an additional population of neurones whose stimulus), or a counterphase moving plaid (a combined Fourier and non-Fourier stimulus), rules out the possibility that the discrepancy between PMAE direction and actual plaid direction is due to the use of test stimuli that do not adequately reflect adaptation by the Fourier and non-Fourier components of the adapting plaids (HR, Ferrera VP, Yo C. Vis Neurosci 1992;9:79-97). Various explanations of this paradoxical result are discussed, including: (i) that MAEs produced by Fourier components out-weigh (and possibly even mask) MAEs produced by non-Fourier plaid components; (ii) PMAEs are influenced by adaptation of a population of component-selective neurones that do not contribute to plaid perception; and, (iii) PMAEs are influenced by component-specific adaptation effects that are weighted according to relative component sensitivity, rather than relative component speed (Pantle A. Vis Res 14;1974:1229-36). We review psychophysical and neurophysiological evidence consistent with these explanations.     

Perception of motion direction in luminance- and contrast-defined reversed-phi motion sequences

Nonlinear processing can be used to recover the motion of contrast modulations of binary noise patterns. A nonlinear stage has also been proposed to explain the perception of forward motion in motion sequences which typically elicit reversed-phi. We examined perceived direction of motion for stimuli in which these reversed motion sequences were used to modulate the contrast of binary noise patterns. A percept of forward motion could be elicted by both luminance-defined and contrast-defined stimuli. The perceived direction of motion seen in the contrast-defined stimuli showed a profound carrier dependency. The replacement of a static carrier by a dynamic carrier can reverse the perceived direction of motion. Forward motion was never seen with dynamic carriers. For luminance- and contrast-defined patterns the reversed motion percept increasingly dominated, with increases in the spatial frequency and temporal frequency of the modulation. Differences in the patterns of responses to the two stimuli over spatial and temporal frequency were abolished by the addition of noise to the luminance-defined stimulus. These data suggest the possibility that a single mechanism may mediate the perception of luminance- and contrast-defined motion.     

Two motion systems with common and separate pathways for color and luminance

We present psychological experiments that reveal two motion systems, a specific and an unspecific one. The specific system prevails at medium to high temporal frequencies. It comprises at least two separate motion pathways that are selective for color and for luminance and that do not interact until after the motion signal is extracted separately in each. By contrast, the unspecific system prevails at low temporal frequencies and it combines color and luminance signals at an earlier stage, before motion extraction. The successful implementation of an efficient and accurate technique for assessing equiluminance corroborates further the main findings. These results offer a general framework for understanding the nature of interactions between color and luminance signals in motion perception and suggest that previously proposed dichotomies in motion processing may be encompassed by the specific/unspecific dichotomy proposed here.     

Binocular correlates of the direction of motion in depth

Two binocular cues to the direction of an object's motion in depth are the ratio (phi R/phi L) between the velocities of the object's retinal images in the right and left eyes and the ratio (phi/gamma) between the velocity of the binocularly-fused image of the object and the rate of change of disparity. We report that the apparent direction of motion in depth of a monocularly-camouflaged target can be varied by altering the ratio phi/gamma. Because no monocular motion signal is available in this case, we conclude that the ratio phi/gamma is a sufficient cue to the direction of motion in depth. This is not to deny that the phi R/phi L cue might be used in the everyday visual situation where monocular velocities phi R and phi L are available to the brain.     

Discrimination of coherent motion when local motion varies in speed and direction.

Random-dot cinematograms (RDCs) consist of multiple local motion signals that can vary in direction and speed. These local motion signals can result in coherent motion: the percept of an overall direction and speed of motion in an RDC. Thresholds were obtained for discriminating differences in the strength of coherent motion. Observers were found to easily discriminate the strength of coherent motion on the basis of the elements' direction or speed under optimal conditions. However, a nonreciprocal relation was evident when this discrimination was performed under nonoptimal conditions. Discrimination of coherent motion that was based on the elements' direction was unaffected, but discrimination that was based on speed was impaired. Results indicate that humans are sensitive to small differences in coherent motion strength and suggest that the visual system processes direction and speed information nonreciprocally. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effects of spatial frequency, duration, and contrast on discriminating motion directions

The minimum speed required for discriminating the direction of drifting gratings was measured at a variety of spatial frequencies, display durations, and contrasts. As was reported previously, speed thresholds were relatively constant for middle and high spatial frequencies, but speed threshold was found to be almost inversely proportional to spatial frequency in the range of 0.25 to 1.0 c/deg. Speed threshold was also found to be inversely proportional to duration between 73 and 40 ms. These results at low frequencies and short durations are shown to be consistent with limits set by the spread of energy in the stimuli, producing velocity uncertainty. A quantitative model of temporal filtering is presented that largely accounts for results at all spatial frequencies and durations by the inclusion of constant positional noise. A discussion includes the possible roles of magnocellular and parvocellular mechanisms in mediating speed thresholds.     

Precise assessment of the mean effective luminance of texture patches--an approach based on reverse-phi motion

In studying the response of mechanisms to contrast-defined texture stimuli, it is critical that the average effective luminance of these textures be equal to that of the background, to minimize net luminance-based signals. We present an efficient and accurate technique for constructing such equiluminant textures to isolate contrast-sensitive mechanisms for investigating their properties. The technique is based on the reverse-phi motion phenomenon, and the resulting settings agree closely with those obtained by photometric means for the class of textures studied. The method also allows one to explore the properties of contrast- and luminance-driven motion mechanisms and, in particular, to evaluate the contribution of putative second-order mechanisms to the motion percept. Results of applying the method are presented, and its advantages over the minimum-flicker and minimum-motion techniques are discussed.     

Similar electrophysiological correlates of texture segregation induced by luminance, orientation, motion and stereo

Certain local features induce preattentive texture segregation. Recently, components in the visual evoked potential (VEP) associated with preattentive texture segregation (tsVEPs) have been demonstrated. To assess the similarity and dissimilarity of visual processing across visual dimensions, we compared VEPs and tsVEPs in texture segregation by luminance, orientation, motion and stereo disparity. We found tsVEPs across these four visual dimensions to be remarkably similar when compared to the "low-level" VEPs. The tsVEPs were always negative; their implicit time, peak latency and amplitude were (in msec/msec/microV): 91/234/-5.7, luminance; 84/257/-3.9, orientation; 80/295/-8.3, motion; and 95/310/-5.0 for stereo. The cross-correlation function, as a quantitative measure for similarity, on average was higher for the tsVEPs by a factor of 4.2 as compared to the low-level VEPs (P < 0.0001). The results suggest (1) that the tsVEPs represent activity of neural mechanisms that have generalised to some degree across visual dimensions; and (2) that these hypothetical generalisation mechanisms might exist already in the primary visual cortex.     

Coherence thresholds for discrimination of motion direction in infants

The sensitivity of 3-month-old infants to direction of motion in random-dot patterns was assessed by measuring coherence thresholds for the discrimination of a pattern, in which opposite directions were segregated into alternate horizontal strips, from an unsegregated pattern. The coherently moving dots had a displacement size of 0.16 deg (velocity 8 deg/sec), and their direction of motion reversed periodically. For both infants and an adult subject coherence thresholds decreased with increasing height of the segregated strips, and with increasing duration of the interval between direction reversals. However the infants required larger minimum heights and longer minimum durations in order to extract motion direction. Even under the best conditions infants were markedly less sensitive, with coherence thresholds of around 50%, compared with 5-7% for the adult. In addition, within the group of infants coherence thresholds were negatively correlated with age. This developmental increase in motion sensitivity at an intermediate velocity suggests that a large part of the improvement in upper and lower velocity thresholds during development is a result of a uniform increase in sensitivity across all velocities, though the results do not rule out additional specific improvements in sensitivity at the extremes of the velocity range.     

Counter-changing luminance: A non-Fourier, nonattentional basis for the perception of single-element apparent motion.

Motion perception was studied for generalized apparent motion stimuli composed of 2 simultaneously visible elements whose luminance alternated between 2 values (only 1 element is visible at a time for standard apparent motion). It was demonstrated that 1st-order motion energy is neither necessary nor sufficient for the perception of apparent motion. Instead, it was found that counter-changing luminance--simultaneous luminance changes at 2 element locations--is the informational basis for perceiving luminance-defined apparent motion: Motion starts where luminance changes toward the background luminance value and ends where luminance changes away from the background luminance. The results were not attributable to either 2nd-order motion mechanisms (for which rectification precedes the computation of motion energy) or attention-based, 3rd-order motion mechanisms. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Adaptive computational models of fast learning of motion direction discrimination

An international conference organised by the European Commission the International Agency of Atomic Energy and the World Health Organization took place on 8-12 April 1996. The conference aimed at evaluating the consequences of the nuclear reactor break down in Chernobyl which have far been identified. The author presents the general outcome of the conference. The article has been prepared on the basis of the conference materials, discussions held during individual sessions and original publications devoted to this subject. All the materials are in the author's possession and can be made available to those interested. The work will be reprinted in extenso in one of the coming issues of the International Journal of Occupational Medicine and Environmental Health.     

The simple reaction time to changes in direction of visual motion

Recently Dzhafarov et al. presented a model explaining data on simple reaction time (RT) to unidimensional velocity changes. The authors suggested that having a motion with an initial velocity V0, the velocity change detection system is reinitialized by means of a "subtractive normalization" process. Therefore, any abrupt change from V0 to V1 is detected as if it were the onset of motion with a speed equal to /V1-V0/. They derived that the RT is a function of /V1-V0/(-2/3). We tested this model for the case of two-dimensional velocity changes. Our subjects observed a random dot pattern that moved horizontally, then changed the direction of motion by an angle alpha in the range between 6 degrees and 180 degrees without changing the speed V. Speeds of 4 and 12 deg/s were used. The subjects reacted as quickly as possible to the direction change. The RTs asymptotically decreased with increasing alpha; with 12 deg/s speed the RTs were shorter than those obtained with 4 deg/s. It was shown that the data can be well described as a function of /V1-V0/(-2/3)=(2Vsin(alpha/2))(-2/3). An extension of the "subtractive normalization" hypothesis for the case of two-dimensional velocity changes is proposed. It is based on the assumption that the velocity vector V1 after the change is decomposed into two orthogonal components. Alternative explanations based on the use of position or orientation cues are shown to contradict the data.