Bistable perception (quantum_mind.families.perception)#
An ambiguous stimulus (a Necker cube, a rotating silhouette, an object a robot’s camera cannot resolve) is perceived as one of two interpretations that alternate. The dwell time in one interpretation is the standard measurement.
Model |
Idea |
Parameters |
|---|---|---|
|
the percept is a qubit that rotates freely between internal observations every Δt; each observation collapses it, so P(switch) = sin²(g Δt) and the mean dwell time ≈ 1/(g² Δt): more frequent observation slows switching (quantum Zeno effect) |
g |
|
classical baseline: constant switching rate, exponential dwell times, no dependence on Δt |
rate |
|
classical baseline: gamma-distributed dwell times, the usual empirical description |
shape, scale |
Distinctive test. With a single observation interval the Zeno and Markov models predict the same geometric or exponential shape. The Zeno model’s own prediction is that the mean dwell time scales as 1/Δt, so a design must vary the observation (presentation or sampling) interval across conditions.
Design: dwell_time_design({condition: dt}) gives histogram bins per condition; dwell_counts(times, edges) turns raw dwell times into counts for fitting (multinomial). For robots, the same models describe
how a perception module that re-samples an ambiguous scene flips between interpretations.
References: Atmanspacher, Filk and Römer, Biological Cybernetics 90, 33–40 (2004); Atmanspacher and Filk, Journal of Mathematical Psychology 54, 314–321 (2010); Levelt, On Binocular Rivalry (1965). Examples use simulated dwell times, labelled as such.