Direction Opponency, Not Quadrature, Is Key to the 1/4 Cycle Preference for Apparent Motion in the Motion Energy Model

Nicolas Heess, Wyeth BairJournal of Neuroscience 2010 · DOI

Abstract

Sensitivity to visual motion is a fundamental property of neurons in the visual cortex and has received wide attention in terms of mathematical models. A key feature of many popular models for cortical motion sensors is the use of pairs of functions that are related by a 90 degrees phase shift. This phase relationship, known as quadrature, is the hallmark of the motion energy model and played an important role in the development of a class of model dubbed elaborated Reichardt detectors. For decades, the literature has supported a link between quadrature and the observation that motion detectors and human observers often prefer a 1/4 cycle displacement of an apparent motion stimulus that consists of a pair of sinusoidal gratings. We show that there is essentially no link between quadrature and this preference. Quadrature is neither necessary nor sufficient for a motion sensor to prefer 1/4 cycle displacement, and motion energy is not maximized for a 1/4 cycle step. Other properties of motion sensors are the key: the opponent subtraction of two oppositely tuned stages that individually have sinusoidal displacement tuning curves. Thus, psychophysical and neurophysiological data revealing a preference at or near 1/4 cycle displacement do not offer specific support for common quadrature or energy-based motion models. Instead, they point to a broader class of model.

Introduction

In visual neuroscience, the "Motion Energy Model" has long served as the gold standard for explaining how the brain perceives movement. Central to this model is the concept of quadrature pairs: two filters that are identical except for a 90° phase shift. Since the mid-1980s, a specific empirical observation—the "1/4 cycle preference"—has been widely accepted as "smoking gun" evidence for these quadrature filters. This preference describes how both human observers and direction-selective neurons often exhibit a peak response when a visual grating stimulus jumps by exactly 1/4 of its spatial cycle (90°).

Heess and Bair (2010) challenge this long-held assumption. They demonstrate that the 1/4 cycle preference is not a unique signature of quadrature-based motion energy units. Instead, they show that this behavior is a fundamental byproduct of direction opponency—the subtraction of signals moving in opposite directions—regardless of the specific phase relationships of the underlying filters.

Methods

The authors re-evaluated the mathematical foundations of motion detection using computational simulations and analytical derivations. They compared three primary architectures:

  1. Motion Energy (ME) Units: Standard models using Gabor-based quadrature pairs.
  2. Reichardt Detectors (RD): Classic correlation-based models that utilize spatial offsets rather than explicit phase shifts.
  3. Generalized Opponent Models: Models using arbitrary (even random) spatiotemporal filters followed by squaring and opponent subtraction.

The study focused on the response of these models to "two-flash" grating stimuli, which are the standard benchmark for measuring displacement tuning in psychophysical and electrophysiological research.

Schematic diagrams of (A) a Motion Energy model using quadrature pairs and (B) an elaborated Reichardt Detector.

Key Insights and Results

Quadrature is Neither Necessary nor Sufficient

A critical finding of the study is that a single quadrature-based motion energy unit does not inherently prefer a 1/4 cycle jump. Its peak response is not fixed at 0.25 cycles; rather, it shifts based on the stimulus's temporal frequency ($f_t$) and the time interval between flashes ($\Delta t$). Analytical derivation shows the optimal displacement $\Delta\phi^*$ is approximately $2\pi f_t \Delta t$.

Displacement tuning curves for a motion energy unit at 5 Hz, 10 Hz, and 20 Hz. The vertical dashed lines indicate that the peak response shifts with frequency and does not remain at the 1/4 cycle (0.25) mark.

Furthermore, the authors show that a Reichardt detector—which lacks quadrature filters—naturally produces a 1/4 cycle preference. This proves that quadrature is not a requirement for this specific stimulus preference.

The Role of Direction Opponency

The researchers discovered that the 1/4 cycle preference is mathematically guaranteed in any system using direction opponency, provided the subunits have sinusoidal displacement tuning. When the response of a sensor tuned to one direction is subtracted from its oppositely-tuned counterpart, the resulting tuning curve follows the form $D^{opp}(\Delta\phi) = 2 \sin(\beta) \sin(\Delta\phi)$. Because the $\sin(\Delta\phi)$ term dominates, the function is forced to peak at $\Delta\phi = \pi/2$, which corresponds exactly to 1/4 cycle. This holds true even for "junk" or random filters, as long as they are paired with an opponent version of themselves.

Conclusion

The 1/4 cycle preference, once thought to be a definitive marker for quadrature filters in the primary visual cortex (V1), is actually a signature of direction opponency. This finding necessitates a re-evaluation of decades of experimental data. It suggests that the 1/4 cycle preference seen in human behavior and neural responses may reflect computations occurring later in the visual hierarchy (such as in area MT/V5), where signals are more robustly opponent, rather than within individual V1 neurons. Future research must look beyond simple displacement tuning to verify the existence of quadrature relationships in the brain.

Summary updated

Abstract

Sensitivity to visual motion is a fundamental property of neurons in the visual cortex and has received wide attention in terms of mathematical models. A key feature of many popular models for cortical motion sensors is the use of pairs of functions that are related by a 90 degrees phase shift. This phase relationship, known as quadrature, is the hallmark of the motion energy model and played an important role in the development of a class of model dubbed elaborated Reichardt detectors. For decades, the literature has supported a link between quadrature and the observation that motion detectors and human observers often prefer a 1/4 cycle displacement of an apparent motion stimulus that consists of a pair of sinusoidal gratings. We show that there is essentially no link between quadrature and this preference. Quadrature is neither necessary nor sufficient for a motion sensor to prefer 1/4 cycle displacement, and motion energy is not maximized for a 1/4 cycle step. Other properties of motion sensors are the key: the opponent subtraction of two oppositely tuned stages that individually have sinusoidal displacement tuning curves. Thus, psychophysical and neurophysiological data revealing a preference at or near 1/4 cycle displacement do not offer specific support for common quadrature or energy-based motion models. Instead, they point to a broader class of model.
BibTeX
@article{heess2010direction,
  author = {Nicolas Heess and Wyeth Bair},
  title = {{Direction Opponency, Not Quadrature, Is Key to the 1/4 Cycle Preference for Apparent Motion in the Motion Energy Model}},
  journal = {Journal of Neuroscience},
  year = {2010},
  doi = {10.1523/jneurosci.1271-10.2010},
}
Figure