[EXPERT: CONSTRUCTED EYE] Day 15 — Luminance Sequences and Motion-Detector Models

The Constructed Eye Masterclass, Day 15
Luminance Sequences and Motion-Detector Models

This sequence of steps and ramps in luminance is used experimentally to probe the responses of motion-detecting circuitry in human and animal visual systems.

Opening Perspective: Artists such as Akiyoshi Kitaoka and perceptual scientists like Beau Lotto have consistently demonstrated and interrogated the fascinating interplay between luminance patterns and our dynamic experience of motion. Day 15 of this masterclass delves into the scientific principles underpinning these effects, focusing on how spatial luminance sequences drive apparent motion and modulate perceptual experience, integrating canonical research models with contemporary art practice.

Expert Objective

Today’s goal is to provide you with a research-level understanding of how luminance sequences—whether gradual gradients or abrupt steps—trigger and modulate motion perception, grounding the discussion in physiological models of motion detection. The session will connect these mechanisms to illusions and studio practice, building on the psychophysical and neural work shaping the field.

A stylized luminance sequence with pronounced steps is commonly used in perceptual experiments on spatial motion direction and the response of the visual cortex.

Evidence and Competing Explanations

• Observed Effects: Apparent motion can be induced by presenting successive static images with spatially shifted luminance sequences (Braddick, 1974; Adelson & Bergen, 1985). Certain luminance sequences produce strong directional motion signals, while others can provoke ambiguous or even reverse apparent motion.
• Supported Mechanisms: Canonical motion-detector models—such as the Reichardt detector (Reichardt, 1961) and elaborations by Adelson & Bergen (1985)—explain how local correlations between spatial and temporal changes in luminance are readout by opponent neural circuitry. These models are directly supported by neurophysiological recordings (Pack & Born, 2001), and echoed in behavioral data. Neurons in the primary visual cortex (V1) and middle temporal area (MT) are known to respond selectively to motion energy defined by specific luminance sequences (Movshon et al., 1985).
• Competing Explanations: Alternative models posit additional roles for higher-order texture and feature tracking, or invoke separate mechanisms for second-order motion (motion defined by contrast or texture, not just luminance). See reviews by Lu & Sperling (2001) and Cavanagh & Mather (1989).
• Unresolved Questions: There remain open questions as to how motion processing integrates ambiguous input, especially when luminance steps are noisy, asymmetric, or manipulated to create reverse-phi effects—where the perceived direction is inverted (Anstis, 1970; Kitaoka graphic illusions).

Qualitative illustration of a reverse-phi motion sequence: when the contrast between steps is inverted between successive frames, the apparent motion direction reverses—a hallmark effect supporting the Reichardt-type opponent process. (After Anstis 1970, Adelson & Bergen 1985)

Digital Experiment

Variables: Use two versions of a spatial sequence: one with simple increasing luminance steps (A → B → C), the other with the contrast of the steps reversed in alternation (A → -B → C).
Protocol: View each sequence in rapid alternation, as fast successive frames (e.g., 60ms per frame). Note the perceived direction of motion: does it follow the real luminance shift, or—when reversing contrast—invert direction (reverse-phi)?
Limitations: This self-experiment is illustrative and demonstrates the effect at the perceptual level, not the underlying neural machinery, which can only be measured by physiological or imaging methods. Individual susceptibility to reverse-phi can vary, especially with fatigue, adaptation, and screen properties.

Retrieval Question

Describe how the motion energy model formalizes the detection of motion in a sequence of luminance steps. What empirical evidence supports that such computations occur in early visual cortex?

Sources

• Adelson EH, Bergen JR. (1985). Spatiotemporal energy models for the perception of motion. Proc Natl Acad Sci USA, 81(17): 5291–5295.
• Pack CC, Born RT. (2001). Temporal dynamics of a neural solution to the aperture problem in visual area MT of macaque brain. Nature, 409(6823): 1040–1042.
• Movshon JA, Adelson EH, Gizzi MS, Newsome WT. (1985). The analysis of moving visual patterns in macaque monkeys. J Neurophysiology, 53(3): 609–635.
• Anstis SM. (1970). Phi movement as a subtraction process. Vision Research, 10(12): 1411-1430.
• Lu ZL, Sperling G. (2001). Three-systems theory of human visual motion perception. Vision Research, 41(20): 2859–2877.
• Cavanagh P, Mather G. (1989). Motion: the long and short of it. Spatial Vision, 4(2-3): 103–129.
• Reichardt, W. (1961). Autocorrelation, a principle for the evaluation of sensory information by the central nervous system. In: Sensory Communication, MIT Press.