The stabilization of visibility for sequentially presented, low-contrast objects: Experiments and neural field model.

In any environment, events transpire in temporal sequences. The general principle governing such sequences is that each instance of the event is influenced by its predecessors. It is shown here that this principle is true for a fundamental aspect of visual perception: visibility. A series of nine psychophysical experiments and associated neural dynamic simulations provide evidence that two non-stimulus factors, self-excitation and short-term memory, stabilize the visibility of a simple low-contrast object (a line segment) as it moves over a sequence of unpredictable locations. Stabilization was indicated by the very low probability of visible-to-invisible switches, and dependence on preceding visibility states was indicated by hysteresis as the contrast of the object was gradually decreased or increased. The contribution of self-excitation to stabilization was indicated by increased visible-to-invisible switching (decreased hysteresis) following adaptation of the visibility state, and the contribution of memory to stabilization was indicated by visibility "bridging" long blank intervals separating each relocation of the object. Because of the unpredictability of the relocations of the object, its visibility at one location pre-shapes visibility at its next location via persisting subthreshold activation of detectors surrounding the low-contrast object. All effects were modeled, including contributions from adaptation and recurrent inhibition, with a single set of parameter values.

Measuring the perceptual grouping of non-adjacent surfaces that are invisibly (amodally) or visibly connected.

Classic Gestalt examples of perceptual grouping entail arrays of disconnected surfaces that are grouped on the basis of the surfaces' relative similarity or proximity. However, most natural environments contain multiple objects, each with multiple, connected surfaces. Moreover, an object in a scene is likely to partially occlude other objects in the 2-dimensional retinal projection of the scene. A central question, therefore, is how the visual system forms a 3-dimensional representation of multi-object scenes by determining which surfaces belong to which objects. To this end, a recently developed dynamic grouping methodology determines whether pairs of surfaces are grouped together on the basis of the direction in which motion is perceived across a surface when its luminance is perturbed. It is shown using this method that the visible surfaces of a partially occluded object are perceptually grouped when they are plausibly connected and represented in a depth plane behind the occluding object. Invisible connectivity (amodal completion) as well as connectivity established by a visible surface have a powerful influence on the grouping of surfaces. However, for neither kind of connectivity is grouping affected by the distance between the surfaces. This absence of a distance/proximity effect on grouping is obtained when the space between to-be-grouped surfaces is filled with other surfaces. It contrasts with the strong effect of distance/proximity on the grouping of disconnected surfaces, and on the clarity of illusory contours formed between disconnected contours. It is concluded that distance/proximity is an operative grouping variable only when there is empty space between the to-be-grouped surfaces.

Temporal Asymmetry in Dark-Bright Processing Initiates Propagating Activity across Primary Visual Cortex.

Differences between visual pathways representing darks and lights have been shown to affect spatial resolution and detection timing. Both psychophysical and physiological studies suggest an underlying retinal origin with amplification in primary visual cortex (V1). Here we show that temporal asymmetries in the processing of darks and lights create motion in terms of propagating activity across V1. Exploiting the high spatiotemporal resolution of voltage-sensitive dye imaging, we captured population responses to abrupt local changes of luminance in cat V1. For stimulation we used two neighboring small squares presented on either bright or dark backgrounds. When a single square changed from dark to bright or vice versa, we found coherent population activity emerging at the respective retinal input locations. However, faster rising and decay times were obtained for the bright to dark than the dark to bright changes. When the two squares changed luminance simultaneously in opposite polarities, we detected a propagating wave front of activity that originated at the cortical location representing the darkened square and rapidly expanded toward the region representing the brightened location. Thus, simultaneous input led to sequential activation across cortical retinotopy. Importantly, this effect was independent of the squares' contrast with the background. We suggest imbalance in dark-bright processing as a driving force in the generation of wave-like activity. Such propagation may convey motion signals and influence perception of shape whenever abrupt shifts in visual objects or gaze cause counterchange of luminance at high-contrast borders. SIGNIFICANCE STATEMENT: An elementary process in vision is the detection of darks and lights through the retina via ON and OFF channels. Psychophysical and physiological studies suggest that differences between these channels affect spatial resolution and detection thresholds. Here we show that temporal asymmetries in the processing of darks and lights create motion signals across visual cortex. Using two neighboring squares, which simultaneously counterchanged luminance, we discovered propagating activity that was strictly drawn out from cortical regions representing the darkened location. Thus, a synchronous stimulus event translated into sequential wave-like brain activation. Such propagation may convey motion signals accessible in higher brain areas, whenever abrupt shifts in visual objects or gaze cause counterchange of luminance at high-contrast borders.

Nonlinear dynamics in the perceptual grouping of connected surfaces.

Evidence obtained using the dynamic grouping method has shown that the grouping of an object's connected surfaces has properties characteristic of a nonlinear dynamical system. When a surface's luminance changes, one of its boundaries is perceived moving across the surface. The direction of this dynamic grouping (DG) motion indicates which of two flanking surfaces has been grouped with the changing surface. A quantitative measure of overall grouping strength (affinity) for adjacent surfaces is provided by the frequency of DG motion perception in directions promoted by the grouping variables. It was found that: (1) variables affecting surface grouping for three-surface objects evolve over time, settling at stable levels within a single fixation, (2) how often DG motion is perceived when a surface's luminance is perturbed (changed) depends on the pre-perturbation affinity state of the surface grouping, (3) grouping variables promoting the same surface grouping combine cooperatively and nonlinearly (super-additively) in determining the surface grouping's affinity, (4) different DG motion directions during different trials indicate that surface grouping can be bistable, which implies that inhibitory interactions have stabilized one of two alternative surface groupings, and (5) when alternative surface groupings have identical affinity, stochastic fluctuations can break the symmetry and inhibitory interactions can then stabilize one of the surface groupings, providing affinity levels are not too high (which results in bidirectional DG motion). A surface-grouping network is proposed within which boundaries vary in salience. Low salience or suppressed boundaries instantiate surface grouping, and DG motion results from changes in boundary salience.

The coupling of vision with locomotion in cortical blindness.

Maintaining or modifying the speed and direction of locomotion requires the coupling of the locomotion with the retinal optic flow that it generates. It is shown that this essential behavioral capability, which requires on-line neural control, is preserved in the cortically blind hemifield of a hemianope. In experiments, optic flow stimuli were presented to either the normal or blind hemifield while the patient was walking on a treadmill. Little difference was found between the hemifields with respect to the coupling (i.e. co-dependency) of optic flow detection with locomotion. Even in the cortically blind hemifield, faster walking resulted in the perceptual slowing of detected optic flow, and self-selected locomotion speeds demonstrated behavioral discrimination between different optic flow speeds. The results indicate that the processing of optic flow, and thereby on-line visuo-locomotor coupling, can take place along neural pathways that function without processing in Area V1, and thus in the absence of conscious intervention. These and earlier findings suggest that optic flow and object motion are processed in parallel along with correlated non-visual locomotion signals. Extrastriate interactions may be responsible for discounting the optical effects of locomotion on the perceived direction of object motion, and maintaining visually guided self-motion.

The independent detection of motion energy and counterchange: flexibility in motion detection.

Motion perception is determined by changing patterns of neural activation initiated by spatiotemporal changes in stimulus features. Motion specified by 1st-order motion energy entails neural patterns that are initiated by spatiotemporal changes in luminance, whereas motion specified by counterchange entails oppositely signed changes in neural activation that can be initiated by spatiotemporal changes in any feature. A constraint in furthering this distinction is that motion energy and counterchange are co-specified by most visual stimuli. In the current study, counterchange was isolated for stimuli composed of translating subjective (Kanizsa) squares, surfaces created by the visual system. Motion energy was isolated for stimuli composed of sequences of luminance increments that spread across perceptually stationary, literal surfaces. Counterchange-specified motion was perceived over a wide range of frame durations, and preferentially for short motion paths. Motion specified by motion energy was diminished for relatively long frame durations, and was unaffected by the length of the motion path. Finally, it was found that blank inter-frame intervals can restore counterchange-specified motion perception for frame durations that are otherwise too brief for motion to be perceived. The results of these and earlier experiments suggest that 1st-order motion energy mechanisms, dedicated to the detection of changes in neural activation initiated by spatiotemporal changes in luminance, provide the basis for objectless motion perception (Wertheimer's phi motion). In contrast, counterchanging neural activation initiated by spatiotemporal changes in any feature, including features created by the visual system, provides a flexible basis for the perception of object motion (Wertheimer's beta motion).

The perception of object versus objectless motion.

Wertheimer, M. (Zeitschrift für Psychologie und Physiologie der Sinnesorgane, 61:161-265, 1912) classical distinction between beta (object) and phi (objectless) motion is elaborated here in a series of experiments concerning competition between two qualitatively different motion percepts, induced by sequential changes in luminance for two-dimensional geometric objects composed of rectangular surfaces. One of these percepts is of spreading-luminance motion that continuously sweeps across the entire object; it exhibits shape invariance and is perceived most strongly for fast speeds. Significantly for the characterization of phi as objectless motion, the spreading luminance does not involve surface boundaries or any other feature; the percept is driven solely by spatiotemporal changes in luminance. Alternatively, and for relatively slow speeds, a discrete series of edge motions can be perceived in the direction opposite to spreading-luminance motion. Akin to beta motion, the edges appear to move through intermediate positions within the object's changing surfaces. Significantly for the characterization of beta as object motion, edge motion exhibits shape dependence and is based on the detection of oppositely signed changes in contrast (i.e., counterchange) for features essential to the determination of an object's shape, the boundaries separating its surfaces. These results are consistent with area MT neurons that differ with respect to speed preference Newsome et al (Journal of Neurophysiology, 55:1340-1351, 1986) and shape dependence Zeki (Journal of Physiology, 236:549-573, 1974).

Changes in pupil diameter entrained by cortically initiated changes in attention.

The diameter of the pupil is affected by changes in ambient illumination, color, spatial structure, movement, and mental effort. It has now been found that pupil diameter can be affected by cognitive processes. That is, it can be entrained by alternations between broadly spread and narrowly focused attention that are cued exogenously (attention is "summoned" by the cue) or endogenously (attention changes under the perceiver's intentional control). Pupil diameter also is affected by post-eye-blink constrictions that occur most often when attention is narrowed, and possibly by changes evoked by the near reflex, although changes in attention state parsimoniously account for the entirety of the results. Changes in pupil diameter produce differences in spherical aberration that alternately blur (when the pupil dilates) and sharpen the retinal image (when the pupil constricts), affecting the relative sensitivity of large receptive fields that mediate broadly spread attention compared with smaller receptive fields that mediate more narrowly focused attention. Results for endogenously cued, intentional changes in attentional spread provide definitive behavioral evidence for cortical feedback to subcortical nuclei that control pupil diameter, either directly or through pupil-constricting eye blinks. Analyses of convergent and divergent changes in eye position indicate that the near reflex was activated long after the initiation of relatively gradual attentionally cued changes in pupil diameter, and further, that when it occurs, the near reflex facilitates ongoing changes in pupil diameter.

Motion perception induced by dynamic grouping: a probe for the compositional structure of objects.

A new method is described for determining how the visual system resolves ambiguities in the compositional structure of multi-surface objects; i.e., how the surfaces of objects are grouped together to form a hierarchical structure. The method entails dynamic grouping motion, a high level process in which changes in a surface (e.g., increases or decreases in its luminance, hue or texture) transiently perturb its affinity with adjacent surfaces. Affinity is determined by the combined effects of Gestalt and other grouping variables in indicating that a pair of surfaces forms a subunit within an object's compositional structure. Such pre-perturbation surface groupings are indicated by the perception of characteristic motions across the changing surface. When the affinity of adjacent surfaces is increased by a dynamic grouping variable, their grouping is transiently strengthened; the perceived motion is away from their boundary. When the affinity of adjacent surfaces is decreased, their grouping is transiently weakened; the perceived motion is toward the surfaces' boundary. It is shown that the affinity of adjacent surfaces depends on the nonlinear, super-additive combination of affinity values ascribable to individual grouping variables, and the effect of dynamic grouping variables on motion perception depends on the prior, pre-perturbation affinity state of the surfaces. It is proposed that affinity-based grouping of an object's surfaces must be consistent with the activation of primitive three-dimensional object components in order for the object to be recognized. Also discussed is the potential use of dynamic grouping for determining the compositional structure of multi-object scenes.

The temporal dynamics of global-to-local feedback in the formation of hierarchical motion patterns: psychophysics and computational simulations.

Four motion quartets, each ambiguous with respect to the perception of parallel-path horizontal or vertical motion, were arranged in a diamond configuration. Both global parallel-path motion (the same motion axis for all the quartets), which is typical for multiquartet stimuli, and global rotational rocking are perceived. Experiment 1 indicated that rotational rocking is established at different levels of processing. Globally, larger displacements of each quartet's elements increase the angle of rotation and, thereby, the perception of rotational rocking. Locally, larger displacements have the opposite effect, weakening motion percepts. Experiment 2 showed that global-to-local feedback affects the local perception of rotation-consistent versus rotation-inconsistent motion directions. Experiment 3 provided evidence for hysteresis effects indicative of competition between global rotational rocking and parallel-path motion. The experimental results were simulated by a two-level dynamical model incorporating global-to-local feedback, with recurrent feedforward/feedback loops creating detection instabilities that amplify activation at both global and local levels of the rotational-rocking pattern.

Illusory motion perception in blindsight.

Motion detection is typically spared in blindsight, which results from damage to the striate cortex (area V1) of the brain that is sufficient to eliminate conscious visual awareness and severely reduce sensitivity to luminance contrast, especially for high spatial and low temporal frequencies. Here we show that the discrimination of motion direction within cortically blind fields is not attributable to feature tracking (the detection of changes in position or shape), but is due instead to the detection of first-order motion energy (spatiotemporal changes in luminance). The key to this finding was a version of the line motion illusion entailing reverse-phi motion in which opposing motion directions are simultaneously cued by motion energy and changes in stimulus shape. In forced-choice tests, a blindsighted test subject selected the direction cued by shape change when the stimulus was presented in his intact field, but reliably selected the direction cued by motion energy when the same stimulus was presented in his blind field, where relevant position information was either inaccessible or invalid. Motion energy has been characterized as objectless, so reliance on motion energy detection is consistent with impaired access to shape information in blindsight. The dissociation of motion direction by visual field (cortically blind vs. intact) provides evidence that two pathways from the retina to MT/V5 (the cortical area specialized for motion perception) are functionally distinct: the retinogeniculate pathway through V1 is specialized for feature-based motion perception, whereas the retinocollicular pathway, which bypasses V1, is specialized for detecting motion energy.

The line motion illusion: the detection of counterchanging edge and surface contrast.

A version of the line motion illusion (LMI) occurs when one of two adjacent surfaces changes in luminance; a new surface is perceived sliding in front of the initially presented surface. Previous research has implicated high-level mechanisms that can create or modulate LMI motion via feedback to lower-level motion detectors. It is shown here that there also is a non-motion-energy, feedforward basis for LMI motion entailing the detection of counterchange, a spatial pattern of motion-specifying stimulus information that combines changes in edge contrast with oppositely signed changes in background-relative surface contrast. It was concluded that (1) in addition to LMI motion, edge/surface counterchange could be the basis for perceiving continuous object motion, (2) counterchange detection is the likely basis for third-order motion perception (Lu & Sperling, 1995a), and (3) motion energy and counterchange mechanisms could be composed of different arrangements of the same spatial and temporal filters, the former detecting motion at a single location, the latter detecting the motion path between pairs of locations.

Measuring perceptual hysteresis with the modified method of limits: dynamics at the threshold.

This article describes modifications to the psychophysical method of limits that eliminate artifacts associated with the classical method, and thereby indicate whether or not there is perceptual hysteresis. Such hysteresis effects, which are characteristic of dynamical systems, would provide evidence that the near-threshold perception of an attribute is affected by stabilization mechanisms intrinsic to individual neural detectors, and by nonlinear interactions that functionally integrate the detectors when there is sufficient stimulus-initiated activation, thereby stabilizing activation at suprathreshold levels. The article begins with a review of research employing the modified method of limits. It concludes with a model and computational simulations showing how detection instabilities inherent in neural dynamics can create 'activational gaps' between the functionally-integrated and functionally-independent states of neural ensembles, resulting in clear and distinct discrimination between the perception and non-perception of an attribute. The 'self-excitation' threshold for engaging such functionally-integrating detector interactions is differentiated from the traditional 'read-out' threshold (criterion) that determines whether or not the attribute in question can be perceived.

Simultaneity and sequence in the perception of apparent motion.

Motion perception usually is accompanied by the phenomenological impression of sequence as objects move through successions of locations. Nonetheless, there is accumulating evidence that sequential information is neither necessary nor sufficient for perceiving motion. It is shown here that apparent motion is specified by counterchange rather than sequence-that is, by co-occurring toward- and away-from-background changes at two spatial locations, regardless of whether the changes are simultaneous or sequential. Motion is perceived from the location of the toward to the location of the away change, even when the changes occur in reverse temporal order. It is not perceived for sequences of away or toward changes, as would be expected if motion were specified by onset or offset asynchronicity. Results previously attributed to onset and offset asynchrony are instead attributable to onsets and offsets occurring in close temporal proximity at the same location. This was consistent with units for detecting away and toward changes that are temporally biphasic; that is, they are excited by changes in one direction and inhibited by immediately preceding or immediately following changes in the opposite direction. These results are accounted for by a model for counterchange-specified motion entailing the biphasic detection of toward and away changes.

A counterchange mechanism for the perception of motion.

A computational model for the perception of counterchange-specified motion is examined in detail and compared with various versions of the Reichardt motion detection model [Reichardt, W. (1961). Autocorrelation, a principle for the evaluation of sensory information by the central nervous system. In W. A. Rosenblith (Ed.), Sensory communication (pp. 303-317). New York: Wiley]. The counterchange model is composed of a pair of temporally biphasic subunits at two retinal locations, one detecting decreases and the other increases in input activation. Motion is signaled when both subunits are simultaneously excited, as determined by the multiplicative combination of their transient responses. In contrast with the Reichardt detector, which effectively tracks motion energy and accounts solely for results obtained with standard apparent motion stimuli (a surface is visible at one location, then at another), the counterchange model also accounts for the generalized apparent motion perceived between pairs of simultaneously visible surfaces. This indicates that standard apparent motion can be perceived via the same non-sequential, non-motion-energy mechanism as generalized apparent motion. There is no need for either an explicit delay mechanism to account for optimal motion perception at non-zero inter-stimulus intervals, or for inhibitory interaction between subunits to account for the absence of motion in the detector's null direction (Barlow, H. B., & Levick, W. R., 1965). Both are emergent properties that result from the inhibitory states of the counterchange detector's biphasic subunits. In addition to apparent motion, the counterchange principle potentially accounts for the perception of motion for drifting gratings, the short range motion perceived for random-dot cinematograms, and the motion perceived for continuously moving objects.

Linking dynamical perceptual decisions at different levels of description in motion pattern formation: psychophysics.

The relationship between local-level motion detection and higher level pattern-forming mechanisms was investigated with the motion quartet, a bistable stimulus for which either horizontal or vertical motion patterns are perceived. Local-level perturbations in luminance contrast affected the stability of the perceived patterns and, thereby, the size of the pattern-level hysteresis obtained by gradually changing the motion quartet's aspect ratio. Briefly eliminating luminance contrast (so nonmotion was perceived during the perturbation) eliminated pattern-level hysteresis, and briefly increasing luminance contrast (so motion was perceived during the perturbation) increased pattern-level hysteresis. Partially reducing luminance contrast resulted in bistability during the perturbation; pattern-level hysteresis was maintained when motion was perceived, and eliminated when nonmotion was perceived. The results were attributed to local motion/nonmotion perceptual decisions in area V1 affecting the magnitude of the activation feeding forward to motion detectors in area MT, where the stability of pattern-level perceptual decisions is determined by activation-dependent, future-shaping interactions that inhibit soon-to-be-stimulated detectors responsive to competing motion directions.

Linking dynamical perceptual decisions at different levels of description in motion pattern formation: computational simulations.

A two-level dynamical model of motion pattern formation is developed in which local motion/ nonmotion perceptual decisions are based on inhibitory competition between area V1 detectors responsive to motion-specifying versus motion-independent stimulus information, and pattern-level perceptual decisions are based on inhibitory competition between area MT motion detectors with orthogonal directional selectivity. The model accounts for the effects of luminance perturbations on the relative size of the pattern-level hysteresis effects reported by Hock and Ploeger (2006) and also accounts for related experimental results reported by Hock, Kelso, and Schöner (1993). Single-trial simulations demonstrated the crucial role of local motion/nonmotion bistability and activation-dependent future-shaping interactions in stabilizing perceived global motion patterns. Such interactions maintain currently perceived motion patterns by inhibiting the soon-to-be-stimulated motion detectors that otherwise would be the basis for the perception of an alternative pattern.

Dynamical vs. judgmental comparison: hysteresis effects in motion perception.

Perceptual comparison was investigated by gradually varying the relative length of two apparent motion paths, and independently determining when an initial percept was lost during the course of attribute change and when an alternative percept emerged. Dynamical comparison was indicated by a range of attribute values for which perception was bistable. Within this range, a percept that lost stability was immediately replaced by an alternative percept. Judgmental comparison was indicated by a range of attribute values for which perception was uncertain. When an initial percept was lost, an alternative percept did not immediately emerge because the alternatives being compared could not be distinguished. Differences in the effects of random noise on dynamical vs. judgmental comparison were demonstrated with computational simulations, and implications are discussed for motion energy models and solutions to the motion correspondence problem.

A common mechanism for the perception of first-order and second-order apparent motion.

A common mechanism for perceiving first-order, luminance-defined, and second-order, texture-contrast defined apparent motion between two element locations is indicated by: (1) transitivity--whether or not motion is perceived is inter-changeably affected by activationally equivalent luminance and contrast changes at each location, (2) local integration--whether or not motion is perceived depends on the net activation change resulting from simultaneous background-relative luminance and background-relative contrast changes at the same element location, and (3) inseparability--apparent motion is not perceived through independent first- or second-order mechanisms when luminance and contrast co-vary at the same location. These results, which are predicted by the response characteristics of directionally selective cells in areas V1, MT, and MST, are not instead attributable to changes in the location of the most salient element (third-order motion), attentive feature tracking, or artifactual first-order motion. Their inconsistency with Lu and Sperling's [Lu, Z., Sperling, G. (1995a). Attention-generated apparent motion. Nature 377, 237, Lu, Z., Sperling, G. (2001). Three-systems theory of human visual motion perception: review and update. Journal of the Optical Society of America A 18, 2331] model, which specifies independent first- and second-order mechanisms, may be due to computational requirements particular to the motion of discrete objects with distinct boundaries defined by spatial differences in luminance, texture contrast, or both.

When motion is not perceived: evidence from adaptation and dynamical stability.

Adaptation was used to probe the perceiver's activation state when either motion or nonmotion percepts are formed for bistable, single-element apparent motion stimuli. Although adaptation was not observed in every instance, when it was observed its effect was to increase the probability of both motion-to-nonmotion and nonmotion-to-motion switches, the time scale of adaptation corresponding to neurophysiological observations for directionally selective cortical cells (Giaschi et al. 1993). This susceptibility to de-stabilizing adaptation effects indicated that the nonmotion percept was not the result of inadequate stimulation producing subthreshold levels of motion detector activation; if that were the case, activation-dependent adaptation would have decreased the nonmotion-to-motion switching rate by reducing activation further below threshold. Above-threshold activation levels are therefore associated with both nonmotion and motion perceptual states, and the failure to perceive motion despite the presence of adequate motion detector stimulation can be attributed to inhibitory competition between detectors activated by motion-specifying stimulus information and detectors activated to similar levels by motion-independent stimulus information, consistent with the dynamical quality of single-element apparent motion.