Motion Perception: How We Detect Movement

Author: Angela Bogdanova · Published: October 16, 2026 · Editorial Policy

Research cluster: Attention, Perception & Cognitive Control · Written by Artificial Sapiens

Motion perception is the visual system's capacity to estimate change in position over time and to interpret the direction, speed, organization, and likely source of that change. It allows a person to follow a moving ball, distinguish a walking person from a shifting background, and maintain a stable sense of the environment while the eyes and body move. Motion perception is an active computation: successive retinal images contain changes in light, and the observer must infer which changes arise from objects, viewpoint movement, illumination, and the structure of the visual scene. A major scientific review describes how motion judgments depend on the interaction of local motion signals with other visual attributes, temporal information, and contextual interpretation Nishida (2011).

The fundamental problem is that the eyes receive a sequence of two-dimensional images rather than direct measurements of the trajectories of three-dimensional objects. A moving shadow, a rotating textured surface, a stationary landscape viewed from a moving car, and two lights flashing in succession can all produce apparent movement, although their physical causes differ. Reliable perception therefore depends on multiple stages of visual processing, the integration of partially ambiguous signals, and the use of scene context. These mechanisms are well established in psychophysics and computational vision; questions about their exact neural implementation, interactions, and generalization across natural environments remain active topics of investigation Adelson and Bergen (1985); Nishida (2011).

What Is Motion Perception?

In experimental psychology, motion perception refers to detecting, discriminating, and interpreting visual displacement and change across time. Detecting motion asks whether there is movement at all; direction discrimination asks whether a pattern travels left rather than right; speed discrimination compares apparent rates of displacement; motion segmentation determines which elements belong to the same moving object; and self-motion perception concerns movement of the observer relative to the environment. These operations can involve overlapping visual computations, but they are distinct measurement targets Nishida (2011).

A stimulus may contain physical displacement without yielding a clear direction judgment, as when a moving pattern is weak or noisy. Conversely, a stationary pattern can seem to move following prolonged exposure to movement, and a succession of stationary images can generate compelling apparent motion. These examples show why motion perception is a perceptual inference rather than a simple readout of object displacement. A visual illusion reveals how a particular computation behaves under specified conditions; it does not establish that everyday motion perception is generally inaccurate Anstis, Verstraten, and Mather (1998); Braddick (1974).

Physical motion, retinal motion, and perceived motion

Physical motion describes change in an object's position within a selected coordinate system. Retinal motion describes image displacement across the photoreceptor array. Perceived motion is the observer's experienced estimate of movement, including speed, direction, object continuity, and sometimes depth or self-motion. When an observer follows a moving target with the eyes, the target may remain near a fixed retinal location although it is moving through the world. Conversely, during an eye movement a stationary room sweeps across the retina without usually appearing to move. A theory of perception must therefore consider eye movements, reference frames, and evidence about the stability of the surrounding scene, rather than equating retinal change with world motion Warren and Hannon (1988); Li (2025).

Why the Visual System Needs Motion Information

Movement provides information about approaching hazards, object trajectories, scene depth, contact with surfaces, and the activity of other organisms. In ordinary vision, motion can help separate objects from a background that shares similar colors or contours. A moving animal may stand out from vegetation even where a still photograph provides only weak segmentation cues. Yet the same stimulus can combine object movement, changes in the viewpoint, and occlusion, requiring the observer to identify what changed and why. Motion is thus part of general perceptual organization, with especially important contributions to navigation and action Nishida (2011); Li (2025).

Understanding this function also requires distinguishing visual motion from action planning. The perception of a moving target supplies estimates that can guide interception, pursuit, and avoidance; successful action additionally depends on motor control, prediction, attention, and environmental constraints. Laboratory measures of motion sensitivity do not themselves measure the whole skill of catching a ball or navigating traffic. Experimental tasks isolate a component of visual processing so that its contribution can be studied with greater precision Britten et al. (1992); Nishida (2011).

How the Brain Detects Visual Movement

A movement signal is distributed across space and time. If a bright edge appears at one location and later at another, a system sensitive to joint spatial and temporal structure can respond differently depending on the apparent direction of travel. Early models used correlations between signals at separated locations or filters tuned to oriented patterns in space-time. In a spatiotemporal representation, a moving feature traces a characteristic orientation; that orientation provides information about velocity. The influential energy model proposed by Adelson and Bergen (1985) showed how pairs of oriented filters can yield direction-sensitive responses without first solving the identity of every object in a sequence.

Direction selectivity is not equivalent to complete motion understanding. A local detector may respond to only a small region, a limited speed range, or a particular spatial frequency. Its response can also change with contrast, texture, temporal frequency, and the angle of an edge. Later stages pool such information and combine it with object boundaries and contextual constraints to estimate the movement of larger patterns. This distinction between local detection and global interpretation is central to classical and current accounts of visual motion Simoncelli and Heeger (1998); Dakin, Mareschal, and Bex (2005).

Spatiotemporal energy and direction-selective responses

A spatiotemporal-energy model represents motion through filters sensitive to particular orientations in the combined dimensions of horizontal position, vertical position, and time. Model responses are squared and pooled in ways that give approximate independence from certain aspects of stimulus phase. Opponent combinations can compare energy signaling alternative directions. This framework predicts important features of motion sensitivity and makes testable predictions about how spatial frequency and temporal frequency shape performance. It is a computational account, rather than a claim that every individual mathematical element maps one-to-one onto a distinct cell Adelson and Bergen (1985).

Local-to-global integration

A scene rarely moves as a set of isolated dots. Individual visible contours may each provide limited information; the visual system has to organize them into a coherent object trajectory. In a field of moving elements, an observer may detect the common direction even when the elements contain substantial variability. Psychophysical work using direction-distributed moving elements separates uncertainty in local estimates from limitations in pooling across elements, demonstrating that both can contribute to a global-motion threshold Dakin, Mareschal, and Bex (2005).

This integration problem becomes harder when different regions genuinely move in different directions, as occurs in a crowd, flowing water, or an object crossing a moving background. Pooling every signal would erase important boundaries. Effective motion interpretation therefore requires both integration within coherent structures and segmentation between independently moving structures. The balance depends on stimulus geometry, occlusion, spatial organization, and temporal duration Nishida (2011).

The Aperture Problem: Why a Moving Edge Can Be Ambiguous

Imagine a diagonal stripe drifting behind a small circular opening. Within that opening, the visual system can measure movement perpendicular to the stripe, but several different real-world trajectories could produce the same local observation. This is the aperture problem. A local motion measurement is underdetermined because the full extent and endpoints of the contour are unavailable. The ambiguity becomes less severe when the observer can use corners, visible line endings, other nearby edges, or the larger configuration of an object Pack and Born (2001).

The aperture problem illustrates why neural and computational motion research distinguishes local vector estimation from pattern-level direction. In macaque area MT, the time course of neural responses to moving patterns suggests that early activity can reflect local component directions and later activity more closely reflects an integrated pattern direction. This result supports a dynamic account of motion integration, while the specific timing and generality depend on stimulus and neural population Pack and Born (2001).

Apparent Motion: Seeing Movement Without Continuous Displacement

Apparent motion occurs when a sequence of spatially separated events is experienced as movement between positions, even if no object travels continuously along the perceived path. Two stationary lights flashing alternately can produce a moving percept when their timing, separation, and other conditions are suitable. The moving percept is not equivalent to a physically moving luminous object, but it demonstrates that the visual system interpolates structure through time. Motion pictures and many animated displays exploit this capacity, although real displays also depend on frame rate, exposure, persistence, tracking, and complex temporal sampling Braddick (1974).

Max Wertheimer's 1912 experiments on apparent movement became foundational for Gestalt psychology because they showed that experienced movement cannot always be derived by treating separately illuminated locations as independent static sensations. Later historical analysis clarified an important distinction: pure phi motion is traditionally distinguished from beta movement, in which an apparent object seems to travel between positions. Many simplified descriptions conflate the two. The scientific lesson concerns the organization of successive events into a moving percept, with distinct phenomenological patterns under different stimulation conditions Steinman, Pizlo, and Pizlo (2000).

Short-range and long-range apparent motion

In classic random-dot experiments, a region can appear to move coherently when its dot pattern is displaced a small amount between presentations. Braddick demonstrated limits on the displacement over which this short-range correspondence supports a segregated moving pattern, suggesting mechanisms different from those underlying some longer-range, more feature-based forms of apparent motion Braddick (1974). Subsequent work complicated any rigid two-channel interpretation: motion can depend on tracking identifiable features, second-order contrast structure, and attention as well as on local luminance-based signals Nishida (2011).

Motion Aftereffects and Adaptation

After watching water flow steadily downward, nearby stationary rocks may appear to drift upward when the observer shifts gaze. This is the familiar waterfall illusion, a motion aftereffect. Exposure to a consistent direction changes the balance of direction-sensitive responses, leading to an illusory signal in the opposite direction when the moving stimulus stops. A review of the phenomenon links it to selective adaptation in motion-sensitive mechanisms at multiple levels of visual processing Anstis, Verstraten, and Mather (1998).

The aftereffect is especially informative because it exposes how prior stimulation affects subsequent perception. Its strength, duration, spatial specificity, and dependence on stimulus characteristics can help distinguish candidate mechanisms. A later reassessment showed that adaptation phenomena include effects beyond a single simple direction-opponent model and that different stimulus conditions reveal different processing levels Mather et al. (2008). An aftereffect in an otherwise ordinary viewing situation is a perceptual phenomenon, not a diagnosis or proof of neurological injury.

Optic Flow: Movement of the Visual World During Self-Motion

When a person walks forward, the retinal image of the surrounding environment changes in an organized way: nearby surfaces generally move faster across the visual field than distant ones, and image trajectories often radiate outward from a region related to the direction of travel. This broad pattern is called optic flow. Unlike the displacement of a single object, optic flow provides information about the movement of the observer through an environment. Its interpretation can help estimate heading, path, scene layout, and the relative motion of independently moving objects Li (2025).

Classic experiments demonstrated that observers can derive direction of self-motion from computer-generated flow fields, including conditions simulating or involving eye movements Warren and Hannon (1988). The result is evidence that optic-flow information can be sufficient under particular experimental conditions. Natural navigation additionally involves body movement, vestibular signals, prior expectations, changing gaze, and environmental structure. Contemporary reviews examine how flow and nonflow information contribute to estimates of current heading and future travel path; conclusions vary with the behavior and display being tested Li (2025).

Object motion versus observer motion

The same image displacement may correspond to a moving object viewed by a stationary person or to a stationary object observed by a moving person. When a car passes another car, both sources of movement may be present. Observers can use common background motion, depth structure, motion discontinuities, and information about eye and body movement to identify the source. An optic-flow field is therefore evidence about self-motion rather than an infallible direct sensor of it. The challenge of independently moving objects remains important in research on navigation through dynamic scenes Li (2025).

Biological Motion: Recognizing Movement of Living Bodies

Humans often recognize a walking person from a sparse arrangement of moving points placed near body joints, despite limited information about the person's full appearance. Biological-motion perception organizes temporal relations among parts into a recognizable movement pattern. Such information can support recognition of actions and aspects of gait; the extent to which it supports judgments about identity or social state depends on what the stimulus preserves and on task conditions. A quantitative framework for analyzing gait patterns showed that characteristic motion information can be decomposed and represented systematically Troje (2002).

Biological motion demonstrates how movement information participates in object and social perception. A constellation of dots can be difficult to interpret while stationary, yet become recognizable as a body in action when its parts move in coordinated ways. Research distinguishes low-level local motion cues, global body configuration, temporal dynamics, and learned expectations. The relative contributions depend on the experiment and remain under study. These findings do not mean that observing a person's gait reliably reveals their intentions, character, or psychiatric status Troje (2002); Nishida (2011).

Speed Perception and Why Context Matters

Perceived speed is influenced by the stimulus as well as by its geometric displacement over time. The contrast, spatial frequency, temporal frequency, viewing duration, eccentricity, and surrounding motion of a pattern can change apparent speed. In a classic experiment with drifting gratings, lowering contrast reduced perceived speed at several relatively low temporal frequencies, whereas the direction of the effect changed under a higher-frequency condition. This interaction is a warning against the simplistic claim that lower contrast always makes motion look slower Thompson (1982).

A physical speed measure and a perceptual judgment should therefore be reported separately. Researchers may ask observers which of two patterns appears faster, estimate a point of subjective equality, or vary a moving stimulus until it reaches a discrimination threshold. These outcomes can be systematically affected by display conditions and reference stimuli. A well-designed experiment identifies the manipulated cue and avoids generalizing a result obtained with laboratory gratings to the perceived speed of every natural object Nishida (2011).

Motion Illusions and Context-Dependent Interpretation

Motion illusions reveal interactions between signals that would otherwise be difficult to isolate. The motion aftereffect shows effects of prior direction-specific exposure; apparent motion shows temporal interpolation across successive images; the aperture problem reveals ambiguities in local contour direction; and contrast-dependent speed judgments show that physically identical displacement can support different estimates. These cases do not all have the same mechanism. Treating them as separate experimental phenomena is more informative than applying one universal explanation of perceptual error Anstis, Verstraten, and Mather (1998); Pack and Born (2001); Thompson (1982).

Context also helps resolve uncertainty. A target may seem to move differently when surrounded by a moving pattern because the visual system estimates relative displacement and partitions the scene into objects and backgrounds. The percept can depend on which region is treated as a stable reference. Motion-based grouping and segmentation are consequently connected to the broader study of perceptual organization, where relations among elements influence what counts as a distinct object Nishida (2011).

What Motion-Selective Brain Areas Contribute

The primate visual system contains interconnected regions with different contributions to motion analysis. Work on area MT, also called V5 in some contexts, has been central because many neurons respond selectively to aspects of stimulus direction and speed. A computational model of the processing relationship between early visual cortex and MT explained patterns of neuronal responses through local filtering, weighted pooling, and normalization. The model serves as a bridge between physiological recordings and computational principles; it does not imply that motion perception is confined to one brain region Simoncelli and Heeger (1998).

Causal evidence came from experiments in macaque monkeys. Targeted disruption of MT selectively elevated thresholds for motion discrimination while leaving one tested contrast-discrimination measure comparatively preserved Newsome and Paré (1988). Related recordings compared individual neurons' sensitivity to weak random-dot motion signals with the animals' own behavioral performance during carefully matched discrimination tasks Britten et al. (1992). These are important converging findings, but they were obtained under specific conditions in a nonhuman primate model. Their interpretation should remain tied to the behavioral tasks and sampled neural systems rather than being generalized to all human visual experiences.

How Scientists Measure Motion Perception

Motion-perception experiments typically manipulate a known property of a visual display and measure responses across trials. A common method is two-alternative forced-choice direction discrimination: an observer reports whether a pattern moves left or right, and performance is measured as motion signal strength changes. In random-dot kinematograms, some dots may follow a coherent direction while others act as noise. The experimenter varies coherence to estimate a performance threshold. This method is particularly useful for studying the relation between signal quality and direction judgments Britten et al. (1992).

Other paradigms use drifting gratings, moving contours, apparent-motion sequences, point-light walkers, optic-flow scenes, or adaptation before a test stimulus. They address different questions and generate different units of measurement. A speed-matching task quantifies apparent velocity; a global-motion task probes integration; an optic-flow heading task asks about the direction of self-movement; a biological-motion task tests recognition of organized bodily movement. Their scores should not be collapsed into a single universal index of visual intelligence or concentration Nishida (2011); Li (2025).

Thresholds, psychometric functions, and reliability

A psychometric function describes how the proportion of a particular response changes with a stimulus variable such as motion coherence or speed difference. Investigators fit a function and identify a specified performance point, such as a conventional threshold. That estimate depends on the task, number of trials, response bias, lapse rate, viewing conditions, and fitting method. For comparisons between groups or across sessions, experimental reliability and measurement invariance require direct attention. A precise laboratory threshold is evidence about the measured task under specified conditions, not automatically a stable personal trait. Classical neural-behavioral comparisons illustrate the rigor necessary for interpretation Britten et al. (1992).

Experimental controls and replication

Researchers must control screen refresh rate, luminance, contrast, physical viewing distance, retinal eccentricity, eye movements, stimulus duration, dot lifetime, and display artifacts. A change in any of these variables can alter the available evidence and sometimes reverse an apparent effect. Preregistration, clearly described analysis decisions, adequate trial numbers, and replication across stimuli improve confidence. Motion research has a substantial base of repeatedly used paradigms and converging experimental methods, but a result from one narrow condition should be stated at its proper level of generality Nishida (2011).

Models and Competing Explanations

Spatiotemporal-energy models describe how local motion-sensitive signals can be computed from spatially and temporally varying intensity. Feature-tracking accounts emphasize correspondence among distinguishable features across images. Pattern-motion integration models explain how ambiguous local signals can produce a coherent direction. Optic-flow approaches focus on the structure of global image movement during self-motion. These approaches answer partly different questions and are frequently complementary rather than mutually exclusive accounts of all movement vision Adelson and Bergen (1985); Simoncelli and Heeger (1998); Li (2025).

Theories differ in what they take as an input, which computations they posit, and how they map onto biological systems. A model may predict behavior without uniquely identifying its neural mechanism; a neural correlate may track a judgment without establishing sole causal responsibility; and a lesion or perturbation can provide causal evidence for a specific task while leaving other motion functions intact. Strong conclusions arise by connecting psychophysical measurements, computational predictions, neural recordings, and selective interventions Newsome and Paré (1988); Pack and Born (2001).

Attention, Expectation, and Motion Processing

Attention can prioritize particular locations, features, or moving objects in a crowded scene. Such selection may improve the use of available sensory evidence, but it is important to separate improvements in detection or decision accuracy from changes in the sensory representation itself. Expectations about where an object will move can guide search and tracking, while abrupt movement can draw attention. Experimental evidence on motion integration also shows why one should not automatically infer a purely attentional explanation from an observed threshold change Nishida (2011).

Motion perception has a specific owner within the psychological encyclopedia because the central question is how movement itself is represented and interpreted. A general discussion of spatial selection belongs to Spatial Attention: The Spotlight of Attention and Cueing Effects; the selection of objects and features is covered by Visual Attention: How We Select Locations, Objects, and Features. Linking these subjects clarifies their relationship while preserving their separate scientific questions.

Development, Variation, and Real-World Use

Sensitivity to visual movement is investigated across the lifespan, but developmental and age-related differences cannot be reduced to a single global motion score. Stimulus parameters, sensory acuity, attention, response demands, and task familiarity can change observed performance. Differences between participants in a controlled laboratory paradigm may also reflect viewing conditions or measurement noise. Any claim about a developmental trajectory or population difference requires evidence from the particular age group and task, ideally with appropriately matched controls. Broad findings in adult psychophysics should not be uncritically transferred to infants, older adults, or clinical populations Nishida (2011).

In practical environments, motion information supports pedestrian navigation, moving-object tracking, and recognition of others' actions, while performance depends on multiple perceptual and motor systems. Driving, athletic skill, and hazard avoidance involve judgments about time, speed, direction, and context, but experimental sensitivity to a drifting grating is not a direct measure of these complex competencies. Similarly, research on training one visual discrimination does not establish broad transferable improvements in concentration or visual safety. Applications must be evaluated in their own settings, with relevant behavioral outcomes Li (2025); Nishida (2011).

Clinical Interpretation and Scientific Boundaries

Motion-perception research is concerned primarily with mechanisms of visual information processing. A low score in a random-dot task, an illusion, or sensitivity to moving patterns is a task-specific observation. It does not by itself constitute a symptom, a screening result with established clinical meaning, or a diagnosis. Clinical evaluation requires additional evidence about persistence, functional impact, medical context, and differential explanations, drawing on appropriately validated assessments. Reports of new or substantial visual disturbance warrant assessment within the relevant eye-care or medical pathway; experimental articles should not substitute for clinical evaluation.

Clinical studies may investigate atypical motion processing in neurological, developmental, or psychiatric populations. Those findings demand separate examination of study quality, comorbid sensory differences, sample recruitment, and diagnostic criteria. The existence of group-average differences cannot establish the status of any individual reader. EN-25 covers the general scientific construct; condition-specific diagnosis and treatment are governed by dedicated clinical subjects, while specialized anatomy and neural mechanisms belong to neighboring neuroscience research Nishida (2011).

Established Findings and Open Research Questions

Several conclusions are well supported: visual movement can be estimated from space-time structure; local direction signals are often ambiguous; pooling and segmentation help construct coherent global movement; sequential static events can elicit apparent motion; prolonged motion can create measurable aftereffects; and flow across the visual field provides information about self-motion. Psychophysical, computational, and neural studies converge on these broad principles, despite differences in experimental scope Adelson and Bergen (1985); Braddick (1974); Pack and Born (2001); Li (2025).

Open questions concern how computations are dynamically coordinated in natural scenes; how object and self-motion are jointly attributed while gaze changes; how feature tracking and energy-based representations interact; how learning and expectations modulate visual estimates; and how models generalize across visual environments, individuals, and recording techniques. Findings remain conditional where stimuli are unusually simplified or sample sizes and methods limit generalization. Increasing experimental realism is valuable when paired with control over the specific mechanisms being tested Nishida (2011); Li (2025).

Frequently Asked Questions

How do we detect motion if the eyes only capture images?

The visual system compares spatially and temporally structured signals and uses mechanisms sensitive to particular motion directions and speeds. Later integration combines local responses into estimates about objects, the background, and self-motion. The result is based on successive images and their relations, rather than on a direct sensory reading of a physical trajectory Adelson and Bergen (1985).

Is motion perception the same as visual attention?

Motion perception concerns detecting and interpreting movement. Visual attention determines which information receives processing priority under competition. Attention can affect motion judgments, and movement can guide attention, but an observed change in attention is not automatically a change in motion sensitivity. Each process has distinct experimental paradigms and explanatory questions Nishida (2011).

Why do movies and flashing lights appear to move?

When stationary images or lights appear in an appropriate temporal and spatial sequence, the perceptual system may connect them into an apparent moving event. The effect depends on presentation conditions and may take different phenomenological forms. Pure phi and beta movement are distinct historical concepts rather than interchangeable labels for every moving display Steinman, Pizlo, and Pizlo (2000).

What is the waterfall illusion?

It is a motion aftereffect: after extended viewing of movement in one direction, a stationary scene may briefly appear to move in the opposite direction. Direction-sensitive adaptation and interactions across visual processing stages contribute to this perceptual phenomenon Anstis, Verstraten, and Mather (1998); Mather et al. (2008).

Why can moving stripes appear to travel in the wrong direction?

A small local view of a long edge cannot fully determine the object's global trajectory. The aperture problem is resolved more effectively as additional features and context become available. Perceived direction also depends on stimulus duration and how local components are combined Pack and Born (2001).

Does motion perception tell us how fast a vehicle is moving?

Visual movement supplies useful estimates of relative motion and the organization of a dynamic scene, but a speed percept is affected by the display and environmental context. Measurements of apparent speed are not identical to an instrument's speed reading, and a single laboratory threshold cannot be substituted for safety-critical real-world judgment Thompson (1982); Li (2025).

Related Articles

Visual Attention: How We Select Locations, Objects, and Features explains how locations, objects, and features are selected when multiple visual events compete for processing.

Spatial Attention: The Spotlight of Attention and Cueing Effects develops the experimental study of location-based selection, cueing, and the distinction between attention and eye position.

Perceptual Load Theory: Why Demanding Tasks Change Distraction analyzes how task demands shape interference and why the results concern attention more directly than motion detection.

References

Adelson, E. H., & Bergen, J. R. (1985). Spatiotemporal energy models for the perception of motion. Journal of the Optical Society of America A, 2(2), 284–299. https://doi.org/10.1364/JOSAA.2.000284

Anstis, S., Verstraten, F. A. J., & Mather, G. (1998). The motion aftereffect. Trends in Cognitive Sciences, 2(3), 111–117. https://doi.org/10.1016/S1364-6613(98)01142-5

Braddick, O. (1974). A short-range process in apparent motion. Vision Research, 14(7), 519–527. https://doi.org/10.1016/0042-6989(74)90041-8

Britten, K. H., Shadlen, M. N., Newsome, W. T., & Movshon, J. A. (1992). The analysis of visual motion: A comparison of neuronal and psychophysical performance. Journal of Neuroscience, 12(12), 4745–4765. https://doi.org/10.1523/JNEUROSCI.12-12-04745.1992

Dakin, S. C., Mareschal, I., & Bex, P. J. (2005). Local and global limitations on direction integration assessed using equivalent noise analysis. Vision Research, 45(24), 3027–3049. https://doi.org/10.1016/j.visres.2005.07.037

Li, L. (2025). Visual perception of self-motion. Annual Review of Vision Science, 11, 447–474. https://doi.org/10.1146/annurev-vision-121423-013200

Mather, G., Pavan, A., Campana, G., & Casco, C. (2008). The motion after-effect reloaded. Trends in Cognitive Sciences, 12(12), 481–487. https://doi.org/10.1016/j.tics.2008.09.002

Newsome, W. T., & Paré, E. B. (1988). A selective impairment of motion perception following lesions of the middle temporal visual area (MT). Journal of Neuroscience, 8(6), 2201–2211. https://doi.org/10.1523/JNEUROSCI.08-06-02201.1988

Nishida, S. (2011). Advancement of motion psychophysics: Review 2001–2010. Journal of Vision, 11(5), Article 11. https://doi.org/10.1167/11.5.11

Pack, C. C., & Born, R. T. (2001). Temporal dynamics of a neural solution to the aperture problem in visual area MT of macaque brain. Nature, 409, 1040–1042. https://doi.org/10.1038/35059085

Simoncelli, E. P., & Heeger, D. J. (1998). A model of neuronal responses in visual area MT. Vision Research, 38(5), 743–761. https://doi.org/10.1016/S0042-6989(97)00183-1

Steinman, R. M., Pizlo, Z., & Pizlo, F. J. (2000). Phi is not beta, and why Wertheimer's discovery launched the Gestalt revolution. Vision Research, 40(17), 2257–2264. https://doi.org/10.1016/S0042-6989(00)00086-9

Thompson, P. (1982). Perceived rate of movement depends on contrast. Vision Research, 22(3), 377–380. https://doi.org/10.1016/0042-6989(82)90153-5

Troje, N. F. (2002). Decomposing biological motion: A framework for analysis and synthesis of human gait patterns. Journal of Vision, 2(5), 371–387. https://doi.org/10.1167/2.5.2

Warren, W. H., Jr., & Hannon, D. J. (1988). Direction of self-motion is perceived from optical flow. Nature, 336, 162–163. https://doi.org/10.1038/336162a0