【Day42 文献精读】A Bayesian Model of Perceived Head-Centered Velocity during Smooth Pursuit Eye Movement

阅读文献：

Freeman, T. C. A., Champion, R. A., & Warren, P. A. (2010). A Bayesian Model of Perceived Head-Centered Velocity during Smooth Pursuit Eye Movement. Current Biology, 20(8), 757-762. https://doi.org/https://doi.org/10.1016/j.cub.2010.02.059 

Summary

1) Aubert Fleishl phenomenon, Filehne illusion, trajectory misperception and the slalom illusion are some illusion that all demonstrate eye speed is underestimated with respect to image speed (Figure 1A).

2) Here we present an alternative Bayesian account, based on a idea that perceptual estimates are increasingly influenced by prior expectations as signals become more uncertain, and on an assumption that the prior for motion is centered on zero (slowness assumption) (Figure 1B).

---

 Results

 Figure 1B:

1) Slowness prior (representing expectations): For motion perception, a plausible prior is that objects are largely at rest. The prior is therefore centered on 0, in which case perceived speed decreases as uncertainty rises.

2) Sensory evidence or signals (likelihood): Sensory evidence is unbiased (i.e., accurate, no literal shift between two distributions of likelihood) but can vary in uncertainty (i.e., precision, dark distribution representing noiser signal is wider than light distribution).

Discrimination trials contained F(Fixation intervals/image motion/relative motion/R)-F or E(Eye pursuit intervals/eye motion/pursuit target motion/T)-E intervals.

Perceived-speed trials contained E-F or F-E intervals.

1) Figure 2B, left column:

    ① The discrimination data in the left column show that thresholds were higher for pursued stimuli(at slow speed), meaning that the speed of pursued stimuli was harder to discriminate than the speed of fixated stimuli→noiser signal→lower certainty or precision

    ② We were able to reproduce the thresholds we found very well, as the lines representing model fit circles representing data well.

      ③ Thresholds were approximately constant for faster speeds when expressed as a proportion of the standard which is known as Weber’s law. But our data also show that Weber’s law breaks down at slow speeds.

2) Figure 2B, right column: For all observers, fixated motion needed to be slowed by around 50% to achieve the perceived-speed match. 

3) Figure 2C: Moreover, the accuracy of eye movements during fixation or pursuit could not explain the data.

Figure 3. Bayesian three-stage model

Velocity v is measured by an unbiased sensory signal with a mean of v. The signal is corrupted by Gaussian internal noise with a mean of 0 and standard deviation given by:

No Filehne illusion (Figure 4A) nor misperception of trajectory could result if sensory measurements were unbiased.

→One way to fix this alternative Bayesian account is to introduce biases into the initial sensory measurements.

---

Discussion

1) Based on the idea that sensory signals encoding the speed of eye motion and image motion differ in precision not accuracy, our Bayesian model is able to explain a range of pursuit-based velocity illusions.

2) It may be that variations in stimulus dimensions like contrast help explain why the degree of underestimation of eye speed varies across the studies shown in Figure 1A.

3) Unlike traditional accounts of head-centered motion perception, our new model emphasizes the role of relative motion and pursuit-target motion.