Speed tuning of motion segmentation and discrimination

Neural encoding of multiple motion speeds in visual cortical area MT

Segmenting objects from each other and their background is critical for vision. The speed at which objects move provides a salient cue for segmentation. However, how the visual system represents and differentiates multiple speeds is largely unknown. Here we investigated the neural encoding of multiple speeds of overlapping stimuli in the primate visual cortex. We first characterized the perceptual capacity of human and monkey subjects to segment spatially overlapping stimuli moving at different speeds. We then determined how neurons in the motion-sensitive, middle-temporal (MT) cortex of macaque monkeys encode multiple speeds. We made a novel finding that the responses of MT neurons to two speeds of overlapping stimuli showed a robust bias toward the faster speed component when both speeds were slow (≤ 20°/s). The faster-speed bias occurred even when a neuron had a slow preferred speed and responded more strongly to the slower component than the faster component when presented alone. The faster-speed bias emerged very early in neuronal response and was robust over time and to manipulations of motion direction and attention. As the stimulus speed increased, the faster-speed bias changed to response averaging. Our finding can be explained by a modified divisive normalization model, in which the weights for the speed components are proportional to the responses of a population of neurons elicited by the individual speeds. Our results suggest that the neuron population, referred to as the weighting pool, includes neurons that have a broad range of speed preferences. As a result, the response weights for the speed components are determined by the stimulus speeds and invariant to the speed preferences of individual neurons. Our findings help to define the neural encoding rule of multiple stimuli and provide new insight into the underlying neural mechanisms. The faster-speed bias would benefit behavioral tasks such as figure-ground segregation if figural objects tend to move faster than the background in the natural environment.

Contribution of the slow motion mechanism to global motion revealed by an MAE technique

Two different motion mechanisms have been identified with motion aftereffect (MAE). (1) A slow motion mechanism, accessed by a static MAE, is sensitive to high-spatial and low-temporal frequency; (2) a fast motion mechanism, accessed by a flicker MAE, is sensitive to low-spatial and high-temporal frequency. We examined their respective responses to global motion after adapting to a global motion pattern constructed of multiple compound Gabor patches arranged circularly. Each compound Gabor patch contained two gratings at different spatial frequencies (0.53 and 2.13 cpd) drifting in opposite directions. The participants reported the direction and duration of the MAE for a variety of global motion patterns. We discovered that static MAE durations depended on the global motion patterns, e.g., longer MAE duration to patches arranged to see rotation than to random motion (Exp 1), and increase with global motion strength (patch number in Exp 2). In contrast, flicker MAEs durations are similar across different patterns and adaptation strength. Further, the global integration occurred at the adaptation stage, rather than at the test stage (Exp 3). These results suggest that slow motion mechanism, assessed by static MAE, integrate motion signals over space while fast motion mechanisms do not, at least under the conditions used.

Speed tuning properties of mirror symmetry detection mechanisms

The human visual system is often tasked with extracting image properties such as symmetry from rapidly moving objects and scenes. The extent to which motion speed and symmetry processing mechanisms interact is not known. Here we examine speed-tuning properties of symmetry detection mechanisms using dynamic dot-patterns containing varying amounts of position and local motion-direction symmetry. We measured symmetry detection thresholds for stimuli in which symmetric and noise elements either drifted with different relative speeds, were relocated at different relative temporal frequencies or were static. We also measured percentage correct responses under two stimulus conditions: a segregated condition in which symmetric and noise elements drifted at different speeds, and a non-segregated condition in which the symmetric elements drifted at two different speeds in equal proportions, as did the noise elements. We found that performance (i) improved gradually with increasing the difference in relative speed between symmetric and noise elements, but was invariant across relative temporal frequencies/lifetime duration differences between symmetric and noise elements, (ii) was higher in the segregated compared to non-segregated conditions, and in the moving compared to the static conditions. We conclude that symmetry detection mechanisms are broadly tuned to speed, with speed-selective symmetry channels combining their outputs by probability summation.

Can speed be judged independent of direction?

The ability to judge speed is a fundamental aspect of visual motion processing. Speed judgments are generally assumed to depend on signals in motion-sensitive, directionally selective, neurons in areas such as V1 and MT. Speed comparisons might therefore be expected to be most accurate when they use information within a common set of directionally tuned neurons. However, there does not appear to be any published evidence on how well speeds can be compared for movements in different directions. We tested speed discrimination judgments between pairs of random-dot stimuli presented side-by-side in a series of four experiments (n = 65). Participants judged which appeared faster of a reference stimulus moving along the cardinal or oblique axis and a comparison stimulus moving either in the same direction or in a different direction. The bias (point of subjective equality) and sensitivity (Weber fraction) were estimated from individual psychometric functions fitted for each condition. There was considerable between-participants variability in psychophysical estimates across conditions. Nonetheless, participants generally made more acute comparisons between stimuli moving in the same direction than those moving in different directions, at least for conditions with an upwards reference (∼20% difference in Weber fractions). We also showed evidence for an oblique effect in speed discrimination when comparing stimuli moving in the same direction, and a bias whereby oblique motion tended to be perceived as moving faster than cardinal motion. These results demonstrate interactions between speed and direction processing, thus informing our understanding of how they are represented in the brain.

Criterion-free measurement of motion transparency perception at different speeds

Transparency perception often occurs when objects within the visual scene partially occlude each other or move at the same time, at different velocities across the same spatial region. Although transparent motion perception has been extensively studied, we still do not understand how the distribution of velocities within a visual scene contribute to transparent perception. Here we use a novel psychophysical procedure to characterize the distribution of velocities in a scene that give rise to transparent motion perception. To prevent participants from adopting a subjective decision criterion when discriminating transparent motion, we used an “odd-one-out,” three-alternative forced-choice procedure. Two intervals contained the standard—a random-dot-kinematogram with dot speeds or directions sampled from a uniform distribution. The other interval contained the comparison—speeds or directions sampled from a distribution with the same range as the standard, but with a notch of different widths removed. Our results suggest that transparent motion perception is driven primarily by relatively slow speeds, and does not emerge when only very fast speeds are present within a visual scene. Transparent perception of moving surfaces is modulated by stimulus-based characteristics, such as the separation between the means of the overlapping distributions or the range of speeds presented within an image. Our work illustrates the utility of using objective, forced-choice methods to reveal the mechanisms underlying motion transparency perception.

Separate requirements for detection and perceptual stability of motion in interocular suppression

In interocular masking, a stimulus presented to one eye (the mask) is made stronger in order to suppress from awareness the target stimulus presented to the other eye. We investigated whether matching the features of the target and the mask would lead to more effective suppression (feature-selective suppression), or not (i.e., non-selective suppression). To control the temporal characteristics of the stimuli, we used a dynamic interocular mask to suppress a moving target, and found that neither matching speed nor pattern of motion led to more effective suppression. Instead, a faster target was detected faster, regardless of the mask type or speed, while a relatively slow (about 1°/s) mask was more perceptually stable (i.e., maintained suppression longer) in a non-selective fashion. While the requirement for target detectability, i.e., salience, is well characterized, relatively little attention is given to the factors that make a mask percept more perceptually stable. Based on these results, we argue that there are separate requirements for detection and perceptual stability.

Probing the Characteristics of Colour-Motion Binding and Its Dependence on Persistent Surface Segregation

Identifying the spatial and temporal characteristics of visual feature binding is a remaining challenge in the science of perception. Within the feature-binding literature, disparate findings have suggested the existence of more than one feature-binding mechanism with differing temporal resolutions. For example, one surprising result is that temporal alternations between two different feature pairings of colour and motion (e.g., orange dots moving left with blue dots moving right) support accurate conjunction discrimination at alternation frequencies of around 10 Hz and greater. However, at lower alternation frequencies around 5 Hz, conjunction discrimination falls to chance. To further investigate this effect, we present two experiments that probe the stimulus characteristics that facilitate or impede feature binding. Using novel manipulations of random dot kinematograms, we identify that facilitating surface representations through temporal integration can enable accurate conjunction discrimination at both intermediate and high alternation frequencies. We also offer a neurally plausible evidence accumulator model to describe these results, removing the need to suggest multiple binding mechanisms acting at different timescales. In effect, we propose a single, flexible binding process, whereby the relatively low temporal resolution for binding features can be circumvented by extracting them from rapidly formed and persistent surface representations.

Reading and coherent motion perception in school age children

This study includes an evaluation, according to age, of the reading and global motion perception developmental trajectories of 2027 school age children in typical stages of development. Reading is assessed using the reading rate score test, for which all of the student participants, regardless of age, received the same passage of text of a medium difficulty reading level. The coherent motion perception threshold is determined according to the adaptive psychophysical protocol based on a four-alternative, forced-choice procedure. Three different dot velocities: 2, 5, and 8 deg/s were used for both assemblies of coherent or randomly moving dots. Reading rate score test results exhibit a wide dispersion across all age groups, so much so that the outlier data overlap, for both the 8 and 18-year-old student-participant age groups. Latvian children's reading fluency developmental trajectories reach maturation at 12-13 years of age. After the age of 13, reading rate scores increase slowly; however, the linear regression slope is different from zero and positive: F(1, 827) = 45.3; p < 0.0001. One hundred eighty-one student-participants having results below the 10th percentile were classified as weak readers in our study group. The reading fluency developmental trajectory of this particular group of student-participants does not exhibit any statistically significant saturation until the age of 18 years old. Coherent motion detection thresholds decrease with age and do not reach saturation. Tests with slower moving dots (2 deg/s) yield results that exhibit significant differences between strong and weak readers.

Linking brain to behavior for the visual perception of figures and objects

The dissociation of a figure from its background is an essential feat of visual perception, as it allows us to detect, recognize, and interact with shapes and objects in our environment. In order to understand how the human brain gives rise to the perception of figures, we here review experiments that explore the links between activity in visual cortex and performance of perceptual tasks related to figure perception. We organize our review according to a proposed model that attempts to contextualize figure processing within the more general framework of object processing in the brain. Overall, the current literature provides us with individual linking hypotheses as to cortical regions that are necessary for particular tasks related to figure perception. Attempts to reach a more complete understanding of how the brain instantiates figure and object perception, however, will have to consider the temporal interaction between the many regions involved, the details of which may vary widely across different tasks.

Perceptual separation of transparent motion components: the interaction of motion, luminance and shape cues

Transparency is perceived when two or more objects or surfaces can be separated by the visual system whilst they are presented in the same region of the visual field at the same time. This segmentation of distinct entities on the basis of overlapping local visual cues poses an interesting challenge for the understanding of cortical information processing. In psychophysical experiments, we studied stimuli that contained randomly positioned disc elements, moving at two different speeds in the same direction, to analyse the interaction of cues during the perception of motion transparency. The current work extends findings from previous experiments with sine wave luminance gratings which only vary in one spatial dimension. The reported experiments manipulate low-level cues, like differences in speed or luminance, and what are likely to be higher level cues such as the relative size of the elements or the superposition rules that govern overlapping regions. The mechanism responsible for separation appears to be mediated by combination of the relevant and available cues. Where perceived transparency is stronger, the neural representations of components are inferred to be more distinguishable from each other across what appear to be multiple cue dimensions. The disproportionally large effect on transparency strength of the type of superposition of disc suggests that with this manipulation, there may be enhanced separation above what might be expected from the linear combination of low-level cues in a process we term labelling. A mechanism for transparency perception consistent with the current results would require a minimum of three stages; in addition to the local motion detection and global pooling and separation of motion signals, findings suggest a powerful additional role of higher level separation cues.

Age-related changes in fine motion direction discriminations

The present study used an equivalent noise method to characterize the sources of reduced performance in fine discrimination of motion with age. We varied the density of the displays, the speed and speed variability and the temporal correlation of dots' motion in successive frames to assess their effect on the sensitivity to motion direction. The results showed that, in all experimental conditions, the older observers had higher levels of internal noise. Both age groups used the stimulus information less efficiently at slow speed and most efficiently when the moving elements were uncorrelated across frames. The older observers were less efficient than the younger observers in all conditions except at high speed where the efficiency of the two age groups was the same. We fitted two biologically plausible models to the experimental data: a modified version of the local-to-global direction encoding model (Dakin et al. in Vis Res 45:3027-3049, 2005) and a model for pooling of motion information in medial temporal area (MT) where the neuronal responses were correlated (Huang and Lisberger in J Neurophysiol 101:3012-3030, 2009). The modeling results indicate that the correlation in neuronal responses is essential to characterize the influence of speed and speed variability on the sensitivity to direction information. For the younger observers, a single set of parameters can account for the effect of noise and the spatio-temporal parameters of the stimuli, while, for the older observers, a change in the correlation of neuronal activity and the directional tuning bandwidth with the levels of external noise is needed. The findings are discussed with respect to optimal use of the dynamic information to overcome the negative effect of aging.

Aging affects the neural representation of speed in Macaque area MT

Human perception of speed declines with age. Much of the decline is probably mediated by changes in the middle temporal (MT) area, an extrastriate area whose neural activity is linked to the perception of speed. In the present study, we used random-dot patterns to study the effects of aging on speed-tuning curves in cortical area MT of macaque visual cortex. Our results provide evidence for a significant degradation of speed selectivity in MT. Cells in old animals preferred lower speeds than did those in young animals. Response modulation and discriminative capacity for speed in old monkeys were also significantly weaker than those in young ones. Concurrently, MT cells in old monkeys showed increased baseline responses, peak responses and response variability, and these changes were accompanied by decreased signal-to-noise ratios. We also found that speed discrimination thresholds in old animals were higher than in young ones. The foregoing neural changes may mediate the declines in visual motion perception that occur during senescence.

The effect of spatial layout on motion segmentation

We present a series of experiments exploring the effect of the stimulus spatial configuration on speed discrimination and two different types of segmentation, for random dot patterns. In the first experiment, we find that parsing the image produces a decrease of speed discrimination thresholds such as was first shown by Verghese and Stone [Verghese, P., & Stone, L. (1997). Spatial layout affects speed discrimination threshold. Vision Research, 37(4), 397-406; Verghese, P., & Stone, L. S. (1996). Perceived visual speed constrained by image segmentation. Nature, 381, 161-163] for sinusoidal gratings. In the second experiment, we study how the spatial configuration affects the ability of a subject in localizing an illusory contour defined by two surfaces with different speeds. Results show that the speed difference necessary to localize the contour decreases as the stimulus patches are separated. The third experiment involves transparency. Our results show a little or null effect for this condition. We explain the first and second experiment in the framework of the model of Bravo and Watamaniuk [Bravo, M., & Watamaniuk, S. (1995). Evidence for two speed signals: a coarse local signal for segregation and a precise global signal for discrimination. Vision Research, 35(12), 1691-1697] who proposed that motion computation consists in, at least, two stages: a first computation of coarse local speeds followed by an integration stage. We propose that the more precise estimate of speed obtained from the integration stage is used to produce a new refined segmentation of the image perhaps, through a feedback loop. Our data suggest that this third stage would not apply to the processing of transparency.

An oblique effect for transparent-motion detection caused by variation in global-motion direction-tuning bandwidths

Despite evidence for the broad direction tuning of global-motion detectors, transparent motion can be detected with comparatively small angular separations. The exact means by which this broad population response is decoded to yield multiple signal directions remains unclear. Consequently, we sought to determine the relationship between angular separation thresholds for transparent motion and the direction-tuning bandwidth of global-motion detectors. Angular separation thresholds were assessed around four axes of motion, with thresholds lower around cardinal axes than the oblique axes. This was also found with lowered signal intensities, despite larger differences between the component directions at threshold, indicating that the transparency oblique effect relies more on the mean direction than the components. Simulations with a model global-motion population suggest this is likely to arise from variation in direction-tuning bandwidths around the cardinal and oblique axes. In a second experiment, adaptation to oblique unidirectional motion produced threshold elevation for a wider range of test directions than adaptation to a cardinal direction. This is consistent with tighter direction tuning around cardinal axes and provides a basis for the transparent-motion oblique effect. Our narrow bandwidth estimates also suggest that transparent-motion detection could rely on bimodal activity within the global-motion stage.

An extension of the transparent-motion detection limit using speed-tuned global-motion systems

When transparent motion is defined purely by direction differences, no more than two signal directions can be detected simultaneously. This limit appears to occur because higher signal intensities are required to detect transparent motion compared with uni-directional motion (Edwards, M., & Greenwood, J. A. (2005). The perception of motion transparency: A signal-to-noise limit. Vision Research, 45, 1877-1884). Increasing the effective signal intensities should therefore increase the number of signals that can be detected. We achieved this by adding speed differences, dividing transparent-motion signals between two speed-tuned global-motion systems. When some signals moved at appropriate low speeds and others at high speeds, up to three signals were detected. This is consistent, at least in part, with the signal-to-noise processing basis of the transparency limit. Differences in contrast polarity were also used to assess whether the limit could be extended using stimulus features without independent global-motion systems. A modest improvement in performance was obtained, suggesting that there may be multiple routes to extending the transparent-motion limit.

The perception of motion transparency: A signal-to-noise limit

A number of studies were conducted to determine how many transparent motion signals observers could simultaneously perceive. It was found that that the limit was two. However, observers required a signal intensity of about 42% in order to perceive a bi-directional transparent stimulus. This signal level was about three times that required to detect a uni-directional motion signal, and higher than was physically possible to achieve in a tri-directional stimulus (in a stimulus in which the different transparent signals are defined only by direction). These results indicate that signal intensity plays an important role in establishing the transparency limit and, as a consequence, implicates the global-motion area (V5/MT) in this process.

Object Motion Computation for the Initiation of Smooth Pursuit Eye Movements in Humans

Pursuing an object with smooth eye movements requires an accurate estimate of its two-dimensional (2D) trajectory. This 2D motion computation requires that different local motion measurements are extracted and combined to recover the global object-motion direction and speed. Several combination rules have been proposed such as vector averaging (VA), intersection of constraints (IOC), or 2D feature tracking (2DFT). To examine this computation, we investigated the time course of smooth pursuit eye movements driven by simple objects of different shapes. For type II diamond (where the direction of true object motion is dramatically different from the vector average of the 1-dimensional edge motions, i.e., VA ≠ IOC = 2DFT), the ocular tracking is initiated in the vector average direction. Over a period of less than 300 ms, the eye-tracking direction converges on the true object motion. The reduction of the tracking error starts before the closing of the oculomotor loop. For type I diamonds (where the direction of true object motion is identical to the vector average direction, i.e., VA = IOC = 2DFT), there is no such bias. We quantified this effect by calculating the direction error between responses to types I and II and measuring its maximum value and time constant. At low contrast and high speeds, the initial bias in tracking direction is larger and takes longer to converge onto the actual object-motion direction. This effect is attenuated with the introduction of more 2D information to the extent that it was totally obliterated with a texture-filled type II diamond. These results suggest a flexible 2D computation for motion integration, which combines all available one-dimensional (edge) and 2D (feature) motion information to refine the estimate of object-motion direction over time.

From 1D to 2D via 3D: dynamics of surface motion segmentation for ocular tracking in primates

In primates, tracking eye movements help vision by stabilising onto the retinas the images of a moving object of interest. This sensorimotor transformation involves several stages of motion processing, from the local measurement of one-dimensional luminance changes up to the integration of first and higher-order local motion cues into a global two-dimensional motion immune to antagonistic motions arising from the surrounding. The dynamics of this surface motion segmentation is reflected into the various components of the tracking responses and its underlying neural mechanisms can be correlated with behaviour at both single-cell and population levels. I review a series of behavioural studies which demonstrate that the neural representation driving eye movements evolves over time from a fast vector average of the outputs of linear and non-linear spatio-temporal filtering to a progressive and slower accurate solution for global motion. Because of the sensitivity of earliest ocular following to binocular disparity, antagonistic visual motion from surfaces located at different depths are filtered out. Thus, global motion integration is restricted within the depth plane of the object to be tracked. Similar dynamics were found at the level of monkey extra-striate areas MT and MST and I suggest that several parallel pathways along the motion stream are involved albeit with different latencies to build-up this accurate surface motion representation. After 200-300 ms, most of the computational problems of early motion processing (aperture problem, motion integration, motion segmentation) are solved and the eye velocity matches the global object velocity to maintain a clear and steady retinal image.

Rapid image-segmentation and perceptual transparency share a process which utilises X-junctions generated by temporal integration in the visual system

Perceptual transparency requires local same-polarity X-junctions, which can also be generated by temporal integration under natural dynamic conditions. In this study, segmentation performance and target appearance were measured for a uniform gray target embedded in a random-dot frame presented with a temporally adjacent mask. Although static cues for both segmentation and transparency were unavailable, transparency was observed only when collinear same-polarity edges reduced backward masking, in both the fovea and the perifovea. These results suggest that the visual system has a common underlying mechanism for rapid segmentation and transparency, which utilises same-polarity X-junctions generated by temporal integration.

The efficiency of speed discrimination for coherent and transparent motion

Transparent motion involves the integration and segmentation of local motion signals. Previous research found a cost for processing transparent random dot motions relative to single coherent motions. However, this cost can be the result of the increased complexity of the transparent stimuli. We investigated this possibility by measuring the efficiency of transparent and coherent motions. Since efficiency normalises human performance to that of an ideal observer in the same task, performance can be compared fairly across tasks. Our task, identical in both transparent and coherent conditions, was to discriminate the fastest speed between two opposite motion directions. In two experiments where we varied dot density and speed, we confirmed the cost in human sensitivity for transparent motion but also found a cost for the ideal observer. The outcome was a consistent residual cost in efficiency for transparent motion. This result points to a processing limitation for transparent motion analogous to previously suggested inhibitory mechanisms between opposite directions of motion. Furthermore, we found that both transparent and coherent motion efficiencies decreased as dot density increased. This latter result stresses the importance of the correspondence problem and suggests that local motion signals are integrated over large areas.

Neural control of on-line guidance of hand reaching movements

Orienting one's gaze towards a peripheral target is usually composed of a hypometric primary saccade followed by a secondary 'corrective saccade' triggered automatically (without conscious perception) by the retinal error at the end of the primary saccade and characterised by a short latency. Due to visual suppression during the saccade, the artificial introduction of a random small target jump during that short period remains undetected and triggers after the end of the primary saccade a normal 'corrective saccade'. As a result this procedure simulates an error in the planning of the primary saccade. On the other hand optimum hand pointing (trade-off between movement time and accuracy) is considered classically to involve a natural parallel initiation of saccade and hand response based on a poor peripheral retinal location, and a further amendment of the hand motor response based on the retinal error provided by the simultaneous vision of target and hand during the movement home phase. To test the hypothesis that the retinal feedback at the end of the primary saccade is used to update the visual target position and amend the ongoing hand motor response, we developed a paradigm involving both an optimum hand pointing and an undetected random target perturbation during the orienting saccade. In order to show that the amendments were controlled by a loop comparing the perceived target location with the dynamic hand position signal, vision of the limb was removed at movement onset. Results showed that the movement was smoothly monitored on-line without additional time processing demands. This functional property of flexibility of the ongoing hand motor response, was generalized from movement extent to movement direction. The undetectability of the perturbation at a conscious level was not a prerequisite for motor flexibility, which was further shown to depend on a critical phase of the limb movement beyond which the latter was no longer amendable, even when the limb was visible. The hand pointing flexibility was further generalised from pointing to the more complex hand reaching and grasping process. It was shown that the flexibility of both the transport and the grasp components were closely coupled. A careful analysis of the data suggested the controlled variable to be the general posture of the upper limb, reaching Bernstein's intuitions about redundancy reduction in skeletomotor systems with degrees of freedom in excess. A kinematics study of the motor flexibility of reaching and grasping in a patient with a bilateral optic ataxia favoured the idea of a posterior parietal cortex involvement in the error processing underlying motor flexibility, reaching the same conclusions as other recent studies using either Positron Emission Tomography or Transcranial Magnetic Stimulation.

The detection of direction-defined and speed-defined spatial contours: one mechanism or two?

It is now accepted that the visual system integrates local orientation information across space to define spatial contours [Vision Research 33 (1993) 173]. More recently, it has been shown that similar integration occurs for the direction of local motion signals, in different parts of the visual field, if they are aligned along the axis of a spatial contour [Vision Research 42 (2002) 653]. Here we ask whether similar spatial-linking rules hold for contours comprised of local elements that share only a common speed (but not direction), in the presence of background elements which collectively have the same mean speed as the contour but considerable random variation in the speeds of the individual elements. Furthermore we investigate the detection of spatial contours that are defined by a common speed that is different (both locally and globally) from that of the background elements. The results show that there is a significant, albeit relatively weak, speed-association field with preferential linking between spatially proximal elements that have similar speeds. Although a salient speed difference between the contour and the background elements enhances detection performance for motion-defined contours, it does so primarily via a different route to that of direction linking. We suggest that for motion-defined contours the Gestalt notions of "common fate" and "good continuity", that describe the parsing of local velocity information into objects, boundaries and contours, are mediated via separate underlying perceptual mechanisms.

Global speed processing: evidence for local averaging within, but not across two speed ranges

A primary task of the visual system is to extract the direction and speed of animate objects from the retinal image. We examined global speed processing by determining how local speeds are integrated and whether integration occurs across all speeds or within fixed speed ranges. The first experiment addressed how local motion signals are combined to determine the speed of an object in motion. Observers judged the speed of a moving cloud of dots that took a random walk in direction while the dots inside the cloud moved somewhat independently of the cloud itself. The apparent speed of the cloud of dots is found to change in proportion with the dot speed and is well predicted by calculating the average speed resulting from nearest neighbour matches across stimulus frames. The second experiment addressed whether local speeds are combined across all speeds or within fixed speed ranges for the detection of global motion. Global dot motion (GDM) stimuli that moved in a radial or rotational directions moving at a low speed of 1.2 degrees /s or a high speed of 9.6 degrees /s were used to measure the thresholds for detecting structured motion as a function of the speed of noise dots (0 degrees /s-10.8 degrees /s) added to the stimulus. With low-speed targets, only additional noise dots moving at low speeds interfered with signal detection. High-speed targets were only interfered with by dots moving at high speeds. This finding established the existence of at least two independent speed tuned systems in the range of speeds tested. Experiment 3 investigated how speed signals are combined within a system to determine the global speed. Using sectored radial GDM stimuli the perceived speed of the fastest dots was measured as a function of whether the speed of the dots in alternate sectors either activated the high or low-speed systems. Averaging only occurred when dots were all within the sensitivity range of the high-speed system, however, if alternate sectors activated separate speed systems, averaging did not occur. Thus local speeds are averaged, independent of direction, to derive a global speed estimate, but averaging only occurs within, and not across, speed tuned mechanisms.

Spatial scale of motion segmentation from speed cues

For the accurate perception of multiple, potentially overlapping, surfaces or objects, the visual system must distinguish different local motion vectors and selectively integrate similar motion vectors over space to segment the retinal image properly. We recently showed that large differences in speed are required to yield a percept of motion transparency. In the present study, to investigate the spatial scale of motion segmentation from speed cues alone, we measured the speed-segmentation threshold (the minimum speed difference required for 75% performance accuracy) for 'corrugated' random-dot patterns, i.e. patterns in which dots with two different speeds were alternately placed in adjacent bars of variable width. In a first experiment, we found that, at large bar widths, a smaller speed difference was required to segment and perceive the corrugated pattern of moving dots, while at small bar-widths, a larger speed difference was required to segment the two speeds and perceive two transparent surfaces of moving dots. Both the perceptual and segmentation performance transitions occurred at a bar width of around 0.4 degrees. In a second experiment, speed-segmentation thresholds were found to increase sharply when dots with different speeds were paired within a local pooling area. The critical pairing distance was about 0.2 degrees in the fovea and increased linearly with stimulus eccentricity. However, across the range of eccentricities tested (up to 15 degrees ), the critical pairing distance did not change much and remained close to the receptive field size of neurons within the primate primary visual cortex. In a third experiment, increasing dot density changed the relationship between speed-segmentation thresholds and bar width. Thresholds decreased for large bar widths, but increased for small bar widths. All of these results are well fit by a simple stochastic model, which estimates the probabilities of having identical or different motion vectors within a local pooling area whose size is the same as that of primate V1 neurons. Altogether, these results demonstrate that speed-based segmentation can function well, even at small spatial scales (i.e. high-spatial frequencies of spatial corrugation) and thereby emphasizes the critical role of a local pooling process early in the cortical motion-processing pathway.

Feedback connections act on the early part of the responses in monkey visual cortex

We previously showed that feedback connections from MT play a role in figure/ground segmentation. Figure/ground coding has been described at the V1 level in the late part of the neuronal responses to visual stimuli, and it has been suggested that these late modulations depend on feedback connections. In the present work we tested whether it actually takes time for this information to be fed back to lower order areas. We analyzed the extracellular responses of 169 V1, V2, and V3 neurons that we recorded in two anesthetized macaque monkeys. MT was inactivated by cooling. We studied the time course of the responses of the neurons that were significantly affected by the inactivation of MT to see whether the effects were delayed relative to the onset of the response. We first measured the time course of the feedback influences from MT on V1, V2, and V3 neurons tested with moving stimuli. For the large majority of the 51 neurons for which the response decreased, the effect was present from the beginning of the response. In the responses averaged after normalization, the decrease of response was significant in the first 10-ms bin of response. A similar result was found for six neurons for which the response significantly increased when MT was inactivated. We then looked at the time course of the responses to flashed stimuli (95 neurons). We observed 15 significant decreases of response and 14 significant increases. In both populations, the effects were significant within the first 10 ms of response. For some neurons with increased responses we even observed a shorter latency when MT was inactivated. We measured the latency of the response to the flashed stimuli. We found that even the earliest responding neurons were affected early by the feedback from MT. This was true for the response to flashed and to moving stimuli. These results show that feedback connections are recruited very early for the treatment of visual information. It further indicates that the presence or absence of feedback effects cannot be deduced from the time course of the response modulations.