[EXPERT: CONSTRUCTED EYE] Day 18 — Checkpoint: Build a Parameterized Peripheral-Drift Generator

Checkpoint: Build a Parameterized Peripheral-Drift Generator

Customizable digital rendering of the classic peripheral drift pattern. When fixating just above or below the bands, lateral motion appears, particularly in the periphery, despite total image stillness.

Day 18 of The Constructed Eye masterclass challenges you to construct a digital peripheral-drift generator, modeling the illusions pioneered by Akiyoshi Kitaoka and neurobiological work by Beau Lotto. We step beyond replication: you’ll parameterize luminance, contrast, and edge geometry to test theoretical models and inform new studio interventions.

Expert Objective

Develop a parameterized graphics tool to systematically manipulate the primary variables implicated in peripheral drift illusions: local luminance skew, contrast edge direction, and spatial frequency. Use this to probe, for yourself and others, the predicted dependencies of illusory motion strength, supporting or challenging competing theoretical accounts from current vision neuroscience literature.

Luminance edge schematic: Gabriel bands, peripheral drift. Peripheral drift illusions depend strongly on edge polarity (light-to-dark vs. dark-to-light) and gradient shape at the patch boundaries. Shown: reverse edge direction to test polarity sensitivity.

Evidence and Competing Explanations

The peripheral drift illusion (PDI), first formalized by Kitaoka (Kitaoka & Ashida, 2003), produces vivid apparent motion in still images, especially in the visual periphery. Though the effect is robust, its mechanisms are debated. Below, we separate key findings, hypotheses, and open questions:

• Observed Effect: When a pattern (e.g. "rotating snakes") contains repeated bands with a slow luminance gradient in one direction and a sharp edge in the other, peripheral viewers report strong motion, aligned with the gradient's shallow-to-steep transition (Conway et al., 2005).
• Supported Mechanisms: Psychophysical and fMRI experiments reveal direction-selective activity in early visual cortex (V1-V3) correlated with gradient polarity (Ashida et al., 2012). Quantitative models suggest that differential temporal response of ON and OFF retinal ganglion pathways to edge slopes leads to temporal asynchrony, driving local motion detectors even for static stimuli (Backus & Oruç, 2005).
• Competing Explanations: Some accounts propose higher-level cognitive filling-in or global context effects, but these lack direct neural correlates or predictive power. Other models attribute the phenomenon to local eye-micro-movements (microsaccades), but controlled studies show PDI strength persists even with stabilized retinal images (Murakami et al., 2006).
• Unresolved Questions: What determines the precise spatial frequency threshold for maximal illusion? Is the effect dominated by retinal nonlinearities or early cortical integration? Can parameterized digital manipulations reliably dissociate these models in complex compositions?

Conceptual diagram: Asynchronous ON/OFF pathway responses to sharp vs. shallow luminance edges. These temporal differences, when fed to direction-selective cortical neurons, can trigger the perception of motion.

Digital Experiment

Purpose: Use the parameterized generator above to test the predicted dependencies on luminance skew and edge direction.

• Manipulated Variables: Luminance gradient direction (polarity), gradient steepness, band width, global contrast.
• Protocol: Fixate a central point just above or below the pattern. Change the gradient steepness and direction; note any difference in perceived speed or direction of drift. Test different display luminances and band widths.
• Controlled Variables: Screen brightness (calibrate for mid-level gray backgrounds), ambient illumination, viewing distance.
• Limitations: Self-observed motion is qualitative; crowd-sourced psychophysical data (e.g. rating scales, forced-choice tests) yield more robust conclusions. Be aware that individual differences and display hardware affect results (Conway et al., 2005).

Retrieval Question

What neural and image-based parameters, supported by current vision research, most strongly modulate the perceived speed and direction of peripheral drift illusions, and how would you manipulate them in a digital system to empirically test between asynchrony and higher-level filling-in accounts?

Sources

• Kitaoka, A., & Ashida, H. (2003). Peripheral drift illusion. Nature Neuroscience.
• Conway, B. R., Kitaoka, A., Yazdanbakhsh, A., Pack, C. C., & Livingstone, M. S. (2005). Neural basis for a powerful static motion illusion. Vision Research.
• Backus, B. T., & Oruç, I. (2005). Illusory motion from change over time in the response to contrast and luminance. Vision Research.
• Ashida, H., Kuriki, I., et al. (2012). Direction-selective neural substrates for the peripheral drift illusion in the human brain. PLoS ONE.
• Murakami, I., Kitaoka, A., & Ashida, H. (2006). A positive correlation between fixation instability and the strength of illusory motion in a static display. Vision Research.