Temporal detection in human vision: dependence on eccentricity.

Studies of human perception of time-varying luminance often aim to estimate either temporal impulse response shapes or temporal modulation transfer functions (MTFs) of putative temporal processing mechanisms. Previously, temporal masking data have been used to estimate the properties and numbers of these temporal mechanisms in central vision for 1 cycle per degree (cpd) targets [Fredericksen and Hess (1998)]. The same methods have been used to explore how these properties change with stimulus energy [Fredericksen and Hess (1997)] and spatial frequency [Fredericksen and Hess (1999)]. We present here analyses of the properties of temporal mechanisms that detect temporal variations of luminance in peripheral vision. The results indicate that a two-filter model provides the best model for our masking data, but that no multiple filter model provides an acceptable fit across the range of parameters varied in this study. Single-filter modelling shows differences between processing mechanisms at 1 cpd in central vision and those that operate eccentrically. We find evidence that this change is because of differences in relative sensitivities of the mechanisms, and to differences in fundamental mechanism impulse responses.

Temporal detection in human vision: dependence on spatial frequency.

In the study of perception of temporal changes in luminance, it is customary to model perceptual performance as based on one or more linear filters. The task is then to estimate the temporal impulse responses or the representation of the impulse response in the frequency domain. Previously, temporal masking data have been used to estimate the properties and numbers of these temporal mechanisms (filters) in central vision for 1-cycle-per-degree (cpd) targets [Vision Res. 38, 1023 (1998)]. The same methods have been used to explore how properties of the estimated filters change with stimulus contrast energy [J. Opt. Soc. Am. A 14, 2557 (1997)]. We present estimated properties for temporal mechanisms that detect low spatial-frequency patterns. The results indicate that two filters provide the best model for performance when mask contrast is significant. There are also differences between properties for mechanisms that detect signal spatial frequencies of 1 cpd and 1/3 cpd. The sensitivity of the low-pass mechanism relative to the bandpass mechanism is reduced at 1/3 cpd, consistent with previous findings.

Integration after adaptation to transparent motion: static and dynamic test patterns result in different aftereffect directions.

One of the many interesting questions in motion aftereffect (MAE) research is concerned with the location(s) along the pathway of visual processing at which certain perceptual manifestations of this illusory motion originate. One such manifestation is the unidirectionality of the MAE after adaptation to moving plaids or transparent motion. This unidirectionality has led to the suggestion that the origin of this MAE might be a single source (gain control) located at, or beyond areas that are believed to be responsible for the integration of motion signals. In this report we present evidence against this suggestion using a simple experiment. For the same adaptation pattern, which consisted of two orthogonally moving transparent patterns with different speeds, we show that the direction of the resulting unidirectional MAE depends on the nature of the test stimulus. We used two kinds of test patterns: static and dynamic. For exactly the same adaptation conditions, the difference in MAE direction between testing with static and dynamic patterns can be as large as 50 degrees. This finding suggests that this MAE is not just a perceptual manifestation of a passive recovery of adapted motion sensors but an active integrative process using the output of different gain controls. A process which takes place after adaptation. These findings are in line with the idea that there are several sites of adaptation along the pathway of visual motion processing and that the nature of the test pattern determines the fate of our perceptual experience of the MAE.

Estimating multiple temporal mechanisms in human vision.

When studying human ability to perceive temporal changes in luminance it is customary to estimate either temporal impulse response shapes or temporal modulation transfer functions, the representation of the impulse response in the frequency domain. The advantages and limitations of previous methods are summarized. We then describe an approach based on use of an impulse response basis set that resolves some of those limitations. We next present psychophysical results for spatiotemporal signal detection in spatiotemporal noise, together with an economical model of performance. The model is based on accepted notions of psychophysical detection mechanisms and the filter basis set described in the first part of the paper. The best-fitting model requires only eight parameters, as opposed to the 198 parameters required to separately fit each psychometric function, and captures both qualitative and quantitative properties of the psychophysical data. Finally, the best-fitting model indicates that only two temporal filters are necessary to describe the performance of each of three subjects under the specific stimulus conditions employed here.

Temporal detection in human vision: dependence on stimulus energy.

We have previously proposed and evaluated an economical model of human performance in tasks requiring spatiotemporal signal detection in spatiotemporal noise [Vision Research (to be published)]. The model was successful in describing human psychophysical performance and provides a means for comparing temporal filters (mechanisms) employed under different stimulus conditions. We present investigations into how estimates of temporal mechanisms depend on the contrast energy of the stimulus. Temporal-sensitivity changes result in covariation of the cutoff and peak frequencies of the low-pass and bandpass mechanisms, respectively, with stimulus energy. The results indicate that sensitivity to high temporal frequencies increases as stimulus energy increases, commensurate with extent physiological evidence in cat and primate.

Independent coding across spatial scales in moving fractal images..

We compared observers' ability to discriminate the direction of apparent motion using images which varied in their spatial characteristic; white or flat spectrum noise, and 1/f noise which has an amplitude spectrum characteristic of natural scenes. The upper spatial limit for discrimination (dmax) was measured using a two-flash random dot kinematogram (RDK), which consisted either of a pair of bandpass filtered images or of a bandpass filtered image and its broadband counterpart. Six bandpass central frequencies were used, ranging from 0.25 to 5.66 cyc/deg. Subjects could perform the direction discrimination task for all six central frequencies in both the bandpass-bandpass and bandpass-broadband sequences for the 1/f images, and dmax values were found to be approximately equal in these two conditions at all spatial scales. However, for the white noise images, direction discrimination was not possible at the lowest central frequencies in the bandpass-broadband task. These data show that information from a wide range of spatial scales is equally salient to the human motion system in images whose amplitude spectra fall as 1/f. However, for white noise images, information at the higher spatial frequencies is more salient and dominates performance in the direction discrimination task. These results are consistent with a model in which spatial frequency filters in the input lines of motion detectors have octave constant spatial frequency bandwidths and equal peak sensitivity. In line with a number of recent studies, this suggests that the spatial properties of motion sensitive cells are matched to the statistical properties of natural scenes.

Responses of complex cells in cat area 17 to apparent motion of random pixel arrays.

The characteristics of directionally selective cells in area 17 of the cat are studied using moving random pixel arrays (RPAs) with 50% white and 50% black pixels. The apparent motion stimulus is similar to that used in human psychophysics [Fredericksen et al. (1993). Vision Research, 33, pp. 1193-1205]. We compare motion sensitivity measured with single-step pixel lifetimes and unlimited pixel lifetimes. A motion stimulus with a single-step pixel lifetime contains directional motion energy primarily at one combination of spatial displacement and temporal delay. We recorded the responses of complex cells to different combinations of displacement and delay to describe their spatio-temporal correlation characteristics. The response to motion of RPAs with unlimited lifetime is strongest along the preferred speed line in a delay vs displacement size diagram. When using an RPA with a single-step pixel lifetime, the cells are responsive to a much smaller range of spatial displacements and temporal delays of the stimulus. The maximum displacement that still gives a directionally selective response is larger when the preferred speed of the cell is higher. It is on average about three times smaller than the receptive field size.

How big is a Gabor patch, and why should we care?

We propose a two-parameter model for the perceived size (spatial extent) of a Gaussian-windowed, drifting sinusoidal luminance pattern (a Gabor patch) based on the simple assumption that perceived size is determined by detection threshold for the sinusoidal carrier. Psychophysical measures of perceived size vary with peak contrast, Gaussian standard deviation, and carrier spatial frequency in a manner predicted by the model. At suprathreshold peak contrasts Gabor perceived size is relatively unaffected by systemic noise but varies in a manner that is consistent with the influence of local contrast gain control. However, at and near threshold, systemic noise plays a major role in determining perceived size. The data and the model indicate that measures of contrast threshold using Gaussian-windowed stimuli (or any other nonflat contrast window) are determined not just by contrast response of the neurons activated by the stimulus but also by integration of that activation over a noisy, contrast-dependent extent of the stimulus in space and time. Thus, when we wish to measure precisely the influence of spatial and temporal integration on threshold, we cannot do so by combining contrast threshold measures with Gaussian-windowed stimuli.

Pitfalls in estimating motion detector receptive field geometry.

A number of psychophysical investigations have used spatial-summation methods to estimate the receptive field (RF) geometry of motion detectors by exploring how psychophysical thresholds change with stimulus height and/or width. This approach is based on the idea that an observer's ability to detect motion direction is strongly determined by the relationship between the stimulus geometry (height and width) and the RF of the activated motion detectors. Our results show that previous estimates of RF geometry can depend significantly on stimulus position in the visual field as well as on the stimulus height-to-width ratio. The data further show that RF estimates depend on the stimulus in a manner that is inconsistent with basic predictions derived from current motion detector models. Hence previous estimates of height, width, and height-to-width ratios of motion detector RFs are inaccurate and unreliable. This inaccuracy/unreliability is attributed to a number of sources. These include incorrect fixed-parameter values in model fits, as well as the confounding of physiological spatial summation area through combined use of contrast thresholds and Gaussian-windowed stimuli. A third source of error is an asymmetric variation of spatiotemporal correlation in the stimulus as either its height or width is varied (and the other dimension held constant). Most importantly, a fourth source of unreliability is attributed to the existence of a nonlinear, nonmonotonic distribution of motion detectors in the visual field that has been previously described and is a natural result of visual anatomy.

Directional motion sensitivity under transparent motion conditions.

We measured directional sensitivity to a foreground pattern while an orthogonally directed background pattern was present under transparent motion conditions. For both foreground and background pattern, the speed was varied between 0.5 and 28 deg sec-1. A multi-step paradigm was employed which results in a better estimation of the suppressive or facilitatory effects than previously applied single-step methods (e.g. measuring Dmax or Dmin). Moreover, our method gives insight into the interactions for a wide range of speed and not just the extreme motion thresholds (the D-values). We found that high background speeds have an inhibitory effect on the detection of a range of high foreground speeds and low background speeds have an inhibitory effect on a range of low foreground speeds. Intermediate background pattern speeds inhibit the detection of both low and high foreground pattern speeds and do so in a systemic manner.

Recovery from adaptation for dynamic and static motion aftereffects: evidence for two mechanisms.

The motion aftereffect (MAE) is an illusory drift of a physically stationary pattern induced by prolonged viewing of a moving pattern. Depending on the nature of the test pattern the MAE can be phenomenally different. This difference in appearance has led to the suggestion that different underlying mechanisms may be responsible and several reports show that this might be the case. Here, we tested whether differences in MAE duration obtained with stationary test patterns and dynamic test patterns can be explained by a single underlying mechanism. We find the results support the existence of (at least) two mechanisms. The two mechanisms show different characteristics: the static MAE (i.e. the MAE tested with a static test pattern) is almost completely stored when the static test is preceded by a dynamic test; in contradistinction, the dynamic MAE is not stored when dynamic testing is preceded by a static test pattern.

An analysis of the temporal integration mechanism in human motion perception.

We present a model for the temporal integration of apparent motion information. The model is constructed by considering psychophysical and neurophysiological data, and consists of the leaky integration of pulsatile motion detector responses to apparent motion stimuli. Each pulse represents a motion detector populational response to a discrete spatial displacement of the spatial pattern. Temporal contrast sensitivity determines the shape of constant-stimulus-duration threshold curves for image frame exposure durations less than about 133 msec. The shape of the threshold curve for image frame exposure durations greater than about 133 msec is determined by the leaky integrator time constant and the shape of the pulses emitted by the motion detectors. The leaky integrator model exhibits threshold saturation behaviour (the reaching of a maximum sensitivity or minimum threshold) seen in psychophysical data as well as dependence of saturation time on the frame rate of the apparent motion stimulus. A low frame rate results in a longer time-to-saturation because the leaky integrator discharges more between detector output pulses. When the motion detector output pulses are far enough apart there is effectively no temporal integration and therefore no threshold improvement over time. Finally, the behaviour of the psychophysical threshold curves across spatial displacement sizes is consistent with a populational-response threshold mechanism combined with spatial summation over a non-uniform distribution of detector types across the visual field.

Spatial summation and its interaction with the temporal integration mechanism in human motion perception.

The combination of visual motion information over visual space (spatial summation) and stimulus duration (temporal integration) was investigated using a random-pixel array (spatiotemporally broad-band) apparent motion stimulus designed to isolate specific populations of visual motion detectors. The results indicate that, in agreement with results from spatiotemporally narrow-band stimuli, spatial summation follows the form of linear probabilistic summation rather than non-linear probabilistic summation. Linear probabilistic summation holds for a wide range of stimulus parameters and when changing either motion stimulus height or width. Linear probabilistic summation breaks down when the motion display region approaches a height and/or width that is related to the spatial displacement size, not the speed, of the random-pixel array. This height and width (termed the critical height and width, or critical dimension), increases with spatial displacement size and can be interpreted as a measure of the basic dimensions of the selected motion detector population's receptive field. The critical height is smaller than the critical width, a result that is consistent with a motion detector receptive field that is elongated in the direction of motion. Perhaps most importantly, the mechanisms of temporal integration and spatial summation can work independently under a wide range of conditions. Finally, the results provide evidence for a short-term inhibitory phenomenon from the edges of the useful display area that affects the visibility of the motion.

Recovery from motion adaptation is delayed by successively presented orthogonal motion.

Following a period of adaptation to a pattern moving in a particular direction, a subsequently viewed stationary pattern appears to move in the opposite direction for some time: the movement after effect (MAE). The MAE lasts longer when the test pattern is not immediately or not continuously presented after adaptation. This phenomenon is called storage. So far research indicates that storage only occurs when textured visual stimulation is absent during part of the test phase or if the processing of a stationary test stimulus is prevented (e.g. by binocular rivalry). We present evidence that storage-like phenomena can occur even while a textured and moving visual stimulus is phenomenally present. We adapted binocularly to uni-directional motion of a random-pixel array M1 for 60 sec. This stimulus was immediately followed by another moving pattern M2. Its motion direction was orthogonal to that of M1. The presentation time of M2 was the independent variable. A stationary pattern was presented immediately after presentation of M2. The direction of the resulting integrated uni-directional MAE was measured. For short presentation times of M2 there is an integrated uni-directional MAE, which shows an interaction of the output of units stimulated by both moving patterns. However, it appeared that the effect of M1 on the direction of this combined uni-directional MAE is much longer present than would be expected from the MAE duration of M1, when tested in isolation.

Movement aftereffect of bi-vectorial transparent motion.

Two moving random-pixel arrays (RPAs) were presented simultaneously in the same target field. These RPAs are perceived as two superimposed transparent moving sheets. Although two directions are perceived simultaneously during stimulus presentation, the movement aftereffect (MAE) is unidirectional. The visual system averages both motion signals in the MAE. For motion vectors of equal magnitude and perpendicular direction the MAE direction is the inverse of the sum of both vectors. In the first experiment we measured perceived direction of the MAE of transparent motion for a range of speed combinations. Results indicate that vector summation only predicts the correct MAE direction for combinations of equal speeds. It is suggested that the direction of the MAE of transparent motion is a resultant of the weighted summation of the component inducing vectors. The question then arises what determines the weighting factors. Directional sensitivity and MAE duration of the individual vectors under transparent conditions were measured and used to weigh the vectors and predict the MAE direction of transparent motion. Statistical analyses showed that MAE duration is a better basis to determine the weighting factors predicting the direction of the MAE of transparent motion than component sensitivity. The direction of the MAE of transparent motion thus seems to be determined by the amount of adaptation to the component vectors as reflected by MAE duration. The results suggest that this gain control cannot be located in the individual motion detectors and must be situated at or after some subsequent cooperation stage of the human motion analysis system.

Temporal integration of random dot apparent motion information in human central vision.

Human motion perception is assumed to be functionally described by an array of bi-local detectors feeding later, higher order computational stages. Using this model as a guide, improvement of spatio-temporal displacement sensitivity by temporal integration (summation) was measured in human central vision using random dot pattern apparent-motion stimuli. Our results agree with previous experiments with regard to improvement of maximum perceivable spatial displacement but show that contrary to previous results the minimum perceivable spatial displacement can be improved in a similar manner. Furthermore, stimulus duration is a more accurate predictor of sensitivity than the number of frames in the stimulus over a wide range of stimulus parameter values. Finally, our results indicate that temporal tuning of motion detectors is inversely related to the size of the spatial pattern displacement.

A transparent motion aftereffect contingent on binocular disparity.

Under transparent motion conditions overlapping surfaces are perceived simultaneously, each with its own direction. The motion aftereffect (MAE) of transparent motion, however, is undirectional and its direction is opposite to that of a sensitivity-weighted vector sum of both inducing vectors. Here we report a bidirectional and transparent MAE contingent on binocular disparity. Depth (from retinal disparity) was introduced between two patterns. A fixation dot was presented at zero disparity, that is, located between the two adaptation patterns. After adaptation to such a stimulus configuration testing was carried out with two stationary test patterns at the same depths as the preceding moving patterns. For opposite directions a clear transparent MAE was perceived. However, if the adaptation directions were orthogonal the chance of a transparent MAE being perceived decreased substantially. This was subject dependent. Some subjects perceived an orthogonal transparent MAE whereas others saw the negative vector sum-an integrated MAE. In addition the behavior of the MAE when the distance in depth between adapting and test patterns was increased was investigated: it was found that the visibility of the MAE then decreased. Visibility is defined in this paper as: (i) the percentage of the trials in which MAEs are perceived and (ii) the average MAE duration. Both measures decreased with increasing distance. The results suggest that segregation and integration may be mediated by direction-tuned channels that interact with disparity-tuned channels.

Spatial integration in coherent motion detection and in the movement aftereffect.

Sensitivity characteristics and spatial integration properties of the motion-detection system are compared with those of the system responsible for the movement aftereffect (MAE), elicited by the same stimulus. This provides new information about the mechanisms involved in MAE generation. A screen was divided into a chequerboard where the squares were filled with random-pixel arrays moving in opposite directions. Changing the size of the squares produced drastic changes in the percept during the adaptation phase and in the MAE during the test phase. One striking new phenomenon that is described is 'structure from MAE.' The results indicate that the receptive fields of units involved in eliciting the MAE are larger than the receptive-field sizes of units involved in detection and segregation of motion components in the stimulus. Furthermore, the results suggest that the receptive fields contributing to the MAE are involved in complex interactions in which different local motion directions are integrated in pattern-specific ways.

Spatio-temporal characteristics of human motion perception.

A bi-local detector array model was assumed to describe the functional performance of monocular motion perception. Distributions of model parameters were measured in human vision at several positions in the visual field. The stimulus paradigm was designed to measure directional motion perception thresholds for individual combinations of spatial displacement and temporal delay in random dot apparent motion stimuli. The resulting data support previous results on perceivable spatial displacement limits in human vision but also indicate that both minimum and maximum perceivable spatial displacement thresholds in human observers have a similar dependence on temporal delay. This dependence changes with eccentricity in the visual field in a qualitatively similar manner but by quantitatively different factors. A description of possible biological properties of the bi-local detector population is presented that may explain how detection of spatio-temporal pattern displacements can be performed by a single system. Such a system also predicts that minimum and maximum perceivable spatial displacement thresholds should scale with visual field eccentricity in a manner consistent with our results.